Merge branch 'clawdbot:main' into main

This commit is contained in:
kastrah 2026-01-18 19:34:43 +01:00 committed by GitHub
commit 5455037eab
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
787 changed files with 45234 additions and 6924 deletions

4
.gitmodules vendored
View File

@ -1,4 +0,0 @@
[submodule "Peekaboo"]
path = Peekaboo
url = https://github.com/steipete/Peekaboo.git
branch = main

View File

@ -1,6 +1,6 @@
# Repository Guidelines
- Repo: https://github.com/clawdbot/clawdbot
- GitHub issues: use literal multiline strings or $'...' for newlines; avoid "\\n" escapes in `gh issue create/edit`.
- GitHub issues/comments/PR comments: use literal multiline strings or `-F - <<'EOF'` (or $'...') for real newlines; never embed "\\n".
## Project Structure & Module Organization
- Source code: `src/` (CLI wiring in `src/cli`, commands in `src/commands`, web provider in `src/provider-web.ts`, infra in `src/infra`, media pipeline in `src/media`).
@ -84,6 +84,7 @@
- When answering questions, respond with high-confidence answers only: verify in code; do not guess.
- Never update the Carbon dependency.
- Any dependency with `pnpm.patchedDependencies` must use an exact version (no `^`/`~`).
- Patching dependencies (pnpm patches, overrides, or vendored changes) requires explicit approval; do not do this by default.
- CLI progress: use `src/cli/progress.ts` (`osc-progress` + `@clack/prompts` spinner); dont hand-roll spinners/bars.
- Status output: keep tables + ANSI-safe wrapping (`src/terminal/table.ts`); `status --all` = read-only/pasteable, `status --deep` = probes.
- Gateway currently runs only as the menubar app; there is no separate LaunchAgent/helper label installed. Restart via the Clawdbot Mac app or `scripts/restart-mac.sh`; to verify/kill use `launchctl print gui/$UID | grep clawdbot` rather than assuming a fixed label. **When debugging on macOS, start/stop the gateway via the app, not ad-hoc tmux sessions; kill any temporary tunnels before handoff.**

View File

@ -2,11 +2,152 @@
Docs: https://docs.clawd.bot
## 2026.1.17-2 (Unreleased)
## 2026.1.18-5
### Changes
- Dependencies: update core + plugin deps (grammy, vitest, openai, Microsoft agents hosting, etc.).
### Fixes
- Configure: hide OpenRouter auto routing model from the model picker. (#1182) — thanks @zerone0x.
## 2026.1.18-4
### Changes
- macOS: switch PeekabooBridge integration to the tagged Swift Package Manager release (no submodule).
- macOS: stop syncing Peekaboo as a git submodule in postinstall.
- Swabble: use the tagged Commander Swift package release.
- CLI: add `clawdbot acp client` interactive ACP harness for debugging.
- Plugins: route command detection/text chunking helpers through the plugin runtime and drop runtime exports from the SDK.
- Plugins: auto-enable bundled channel/provider plugins when configuration is present.
- Config: stamp last-touched metadata on write and warn if the config is newer than the running build.
- macOS: hide usage section when usage is unavailable instead of showing provider errors.
- Memory: add native Gemini embeddings provider for memory search. (#1151)
- Agents: add local docs path resolution and include docs/mirror/source/community pointers in the system prompt.
- Slack: add HTTP webhook mode via Bolt HTTP receiver for Events API deployments. (#1143) — thanks @jdrhyne.
### Fixes
- Auth profiles: keep auto-pinned preference while allowing rotation on failover; user pins stay locked. (#1138) — thanks @cheeeee.
- Agents: sanitize oversized image payloads before send and surface image-dimension errors.
- macOS: Doctor repairs LaunchAgent bootstrap issues for Gateway + Node when listed but not loaded. (#1166) — thanks @AlexMikhalev.
- macOS: avoid touching launchd in Remote over SSH so quitting the app no longer disables the remote gateway. (#1105)
- Memory: index atomically so failed reindex preserves the previous memory database. (#1151)
- Memory: avoid sqlite-vec unique constraint failures when reindexing duplicate chunk ids. (#1151)
## 2026.1.18-3
### Changes
- Exec: add host/security/ask routing for gateway + node exec.
- Exec: add `/exec` directive for per-session exec defaults (host/security/ask/node).
- macOS: migrate exec approvals to `~/.clawdbot/exec-approvals.json` with per-agent allowlists and skill auto-allow toggle.
- macOS: add approvals socket UI server + node exec lifecycle events.
- Nodes: add headless node host (`clawdbot node start`) for `system.run`/`system.which`.
- Nodes: add node daemon service install/status/start/stop/restart.
- Bridge: add `skills.bins` RPC to support node host auto-allow skill bins.
- Slash commands: replace `/cost` with `/usage off|tokens|full` to control per-response usage footer; `/usage` no longer aliases `/status`. (Supersedes #1140) — thanks @Nachx639.
- Sessions: add daily reset policy with per-type overrides and idle windows (default 4am local), preserving legacy idle-only configs. (#1146) — thanks @austinm911.
- Agents: auto-inject local image references for vision models and avoid reloading history images. (#1098) — thanks @tyler6204.
- Docs: refresh exec/elevated/exec-approvals docs for the new flow. https://docs.clawd.bot/tools/exec-approvals
- Docs: add node host CLI + update exec approvals/bridge protocol docs. https://docs.clawd.bot/cli/node
- ACP: add experimental ACP support for IDE integrations (`clawdbot acp`). Thanks @visionik.
- Tools: allow `sessions_spawn` to override thinking level for sub-agent runs.
- Channels: unify thread/topic allowlist matching + command/mention gating helpers across core providers.
- Models: add Qwen Portal OAuth provider support. (#1120) — thanks @mukhtharcm.
- Memory: add `--verbose` logging for memory status + batch indexing details.
- Memory: allow parallel OpenAI batch indexing jobs (default concurrency: 2).
- macOS: add per-agent exec approvals with allowlists, skill CLI auto-allow, and settings UI.
- Docs: add exec approvals guide and link from tools index. https://docs.clawd.bot/tools/exec-approvals
- macOS: add exec-host IPC for node service `system.run` with HMAC + peer UID checks.
### Fixes
- Exec approvals: enforce allowlist when ask is off; prefer raw command for node approvals/events.
- Tools: return a companion-app-required message when node exec is requested with no paired node.
- Streaming: emit assistant deltas for OpenAI-compatible SSE chunks. (#1147) — thanks @alauppe.
- Model fallback: treat timeout aborts as failover while preserving user aborts. (#1137) — thanks @cheeeee.
## 2026.1.18-2
### Fixes
- Tests: stabilize plugin SDK resolution and embedded agent timeouts.
## 2026.1.18-1
### Changes
- Tools: allow `sessions_spawn` to override thinking level for sub-agent runs.
- Channels: unify thread/topic allowlist matching + command/mention gating helpers across core providers.
- Models: add Qwen Portal OAuth provider support. (#1120) — thanks @mukhtharcm.
- Memory: add `--verbose` logging for memory status + batch indexing details.
- Memory: allow parallel OpenAI batch indexing jobs (default concurrency: 2).
- macOS: add per-agent exec approvals with allowlists, skill CLI auto-allow, and settings UI.
- Docs: add exec approvals guide and link from tools index. https://docs.clawd.bot/tools/exec-approvals
### Fixes
- Memory: apply OpenAI batch defaults even without explicit remote config.
- macOS: bundle Textual resources in packaged app builds to avoid code block crashes. (#1006)
- Tools: return a companion-app-required message when `system.run` is requested without a supporting node.
- Discord: only emit slow listener warnings after 30s.
## 2026.1.17-6
### Changes
- Plugins: add exclusive plugin slots with a dedicated memory slot selector.
- Memory: ship core memory tools + CLI as the bundled `memory-core` plugin.
- Docs: document plugin slots and memory plugin behavior.
- Plugins: add the bundled BlueBubbles channel plugin (disabled by default).
- Plugins: migrate bundled messaging extensions to the plugin SDK; resolve plugin-sdk imports in loader.
- Plugins: migrate the Zalo plugin to the shared plugin SDK runtime.
- Plugins: migrate the Zalo Personal plugin to the shared plugin SDK runtime.
## 2026.1.17-5
### Changes
- Memory: add hybrid BM25 + vector search (FTS5) with weighted merging and fallback.
- Memory: add SQLite embedding cache to speed up reindexing and frequent updates.
- CLI: surface FTS + embedding cache state in `clawdbot memory status`.
- Memory: render progress immediately, color batch statuses in verbose logs, and poll OpenAI batch status every 2s by default.
- Plugins: allow optional agent tools with explicit allowlists and add plugin tool authoring guide. https://docs.clawd.bot/plugins/agent-tools
- Tools: centralize plugin tool policy helpers.
- Commands: add `/subagents info` and show sub-agent counts in `/status`.
- Docs: clarify plugin agent tool configuration. https://docs.clawd.bot/plugins/agent-tools
### Fixes
- Voice call: include request query in Twilio webhook verification when publicUrl is set. (#864)
## 2026.1.18-1
### Changes
- Tools: allow `sessions_spawn` to override thinking level for sub-agent runs.
- Channels: unify thread/topic allowlist matching + command/mention gating helpers across core providers.
- Models: add Qwen Portal OAuth provider support. (#1120) — thanks @mukhtharcm.
- Memory: add `--verbose` logging for memory status + batch indexing details.
- Memory: allow parallel OpenAI batch indexing jobs (default concurrency: 2).
- macOS: add per-agent exec approvals with allowlists, skill CLI auto-allow, and settings UI.
- Docs: add exec approvals guide and link from tools index. https://docs.clawd.bot/tools/exec-approvals
### Fixes
- Memory: apply OpenAI batch defaults even without explicit remote config.
- macOS: bundle Textual resources in packaged app builds to avoid code block crashes. (#1006)
- Tools: return a companion-app-required message when `system.run` is requested without a supporting node.
- Discord: only emit slow listener warnings after 30s.
## 2026.1.17-3
### Changes
- Memory: add OpenAI Batch API indexing for embeddings when configured.
- Memory: enable OpenAI batch indexing by default for OpenAI embeddings.
### Fixes
- Memory: retry transient 5xx errors (Cloudflare) during embedding indexing.
## 2026.1.17-2
### Changes
### Fixes
- Tools: show exec elevated flag before the command and keep it outside markdown in tool summaries.
- Memory: parallelize embedding indexing with rate-limit retries.
- Memory: split overly long lines to keep embeddings under token limits.
- Memory: skip empty chunks to avoid invalid embedding inputs.
- Sessions: fall back to session labels when listing display names. (#1124) — thanks @abdaraxus.
- Discord: inherit parent channel allowlists for thread slash commands and reactions. (#1123) — thanks @thewilloftheshadow.
## 2026.1.17-1
@ -32,6 +173,7 @@ Docs: https://docs.clawd.bot
### Changes
- CLI: stamp build commit into dist metadata so banners show the commit in npm installs.
- CLI: close memory manager after memory commands to avoid hanging processes. (#1127) — thanks @NicholasSpisak.
## 2026.1.16-1

@ -1 +0,0 @@
Subproject commit 5c195f5e46ebfcc953af74fdd05fbc962d05a50c

View File

@ -249,7 +249,7 @@ Send these in WhatsApp/Telegram/Slack/Microsoft Teams/WebChat (group commands ar
- `/compact` — compact session context (summary)
- `/think <level>` — off|minimal|low|medium|high|xhigh (GPT-5.2 + Codex models only)
- `/verbose on|off`
- `/cost on|off` — append per-response token/cost usage lines
- `/usage off|tokens|full` — per-response usage footer
- `/restart` — restart the gateway (owner-only in groups)
- `/activation mention|always` — group activation toggle (groups only)

View File

@ -13,7 +13,7 @@ let package = Package(
.executable(name: "swabble", targets: ["SwabbleCLI"]),
],
dependencies: [
.package(path: "../Peekaboo/Commander"),
.package(url: "https://github.com/steipete/Commander.git", exact: "0.2.1"),
.package(url: "https://github.com/apple/swift-testing", from: "0.99.0"),
],
targets: [

View File

@ -1,6 +1,24 @@
{
"originHash" : "7eec77e2b399c480e76fdfc7dc3162652f5c775530e9fc282953de38ef2de79b",
"originHash" : "4ed05a95fa9feada29b97f81b3194392e59a0c7b9edf24851f922bc2b72b0438",
"pins" : [
{
"identity" : "axorcist",
"kind" : "remoteSourceControl",
"location" : "https://github.com/steipete/AXorcist.git",
"state" : {
"revision" : "c75d06f7f93e264a9786edc2b78c04973061cb2f",
"version" : "0.1.0"
}
},
{
"identity" : "commander",
"kind" : "remoteSourceControl",
"location" : "https://github.com/steipete/Commander.git",
"state" : {
"revision" : "9e349575c8e3c6745e81fe19e5bb5efa01b078ce",
"version" : "0.2.1"
}
},
{
"identity" : "elevenlabskit",
"kind" : "remoteSourceControl",
@ -10,15 +28,6 @@
"version" : "0.1.0"
}
},
{
"identity" : "eventsource",
"kind" : "remoteSourceControl",
"location" : "https://github.com/mattt/eventsource.git",
"state" : {
"revision" : "ca2a9d90cbe49e09b92f4b6ebd922c03ebea51d0",
"version" : "1.3.0"
}
},
{
"identity" : "menubarextraaccess",
"kind" : "remoteSourceControl",
@ -28,6 +37,15 @@
"version" : "1.2.2"
}
},
{
"identity" : "peekaboo",
"kind" : "remoteSourceControl",
"location" : "https://github.com/steipete/Peekaboo.git",
"state" : {
"branch" : "main",
"revision" : "bace59f90bb276f1c6fb613acfda3935ec4a7a90"
}
},
{
"identity" : "sparkle",
"kind" : "remoteSourceControl",
@ -46,33 +64,6 @@
"version" : "1.2.1"
}
},
{
"identity" : "swift-asn1",
"kind" : "remoteSourceControl",
"location" : "https://github.com/apple/swift-asn1.git",
"state" : {
"revision" : "810496cf121e525d660cd0ea89a758740476b85f",
"version" : "1.5.1"
}
},
{
"identity" : "swift-async-algorithms",
"kind" : "remoteSourceControl",
"location" : "https://github.com/apple/swift-async-algorithms",
"state" : {
"revision" : "6c050d5ef8e1aa6342528460db614e9770d7f804",
"version" : "1.1.1"
}
},
{
"identity" : "swift-collections",
"kind" : "remoteSourceControl",
"location" : "https://github.com/apple/swift-collections",
"state" : {
"branch" : "main",
"revision" : "8e5e4a8f3617283b556064574651fc0869943c9a"
}
},
{
"identity" : "swift-concurrency-extras",
"kind" : "remoteSourceControl",
@ -82,24 +73,6 @@
"version" : "1.3.2"
}
},
{
"identity" : "swift-configuration",
"kind" : "remoteSourceControl",
"location" : "https://github.com/apple/swift-configuration",
"state" : {
"revision" : "3528deb75256d7dcbb0d71fa75077caae0a8c749",
"version" : "1.0.0"
}
},
{
"identity" : "swift-crypto",
"kind" : "remoteSourceControl",
"location" : "https://github.com/apple/swift-crypto.git",
"state" : {
"revision" : "6f70fa9eab24c1fd982af18c281c4525d05e3095",
"version" : "4.2.0"
}
},
{
"identity" : "swift-log",
"kind" : "remoteSourceControl",
@ -118,24 +91,6 @@
"version" : "1.1.1"
}
},
{
"identity" : "swift-sdk",
"kind" : "remoteSourceControl",
"location" : "https://github.com/modelcontextprotocol/swift-sdk.git",
"state" : {
"revision" : "c0407a0b52677cb395d824cac2879b963075ba8c",
"version" : "0.10.2"
}
},
{
"identity" : "swift-service-lifecycle",
"kind" : "remoteSourceControl",
"location" : "https://github.com/swift-server/swift-service-lifecycle",
"state" : {
"revision" : "1de37290c0ab3c5a96028e0f02911b672fd42348",
"version" : "2.9.1"
}
},
{
"identity" : "swift-subprocess",
"kind" : "remoteSourceControl",

View File

@ -20,10 +20,9 @@ let package = Package(
.package(url: "https://github.com/swiftlang/swift-subprocess.git", from: "0.1.0"),
.package(url: "https://github.com/apple/swift-log.git", from: "1.8.0"),
.package(url: "https://github.com/sparkle-project/Sparkle", from: "2.8.1"),
.package(url: "https://github.com/steipete/Peekaboo.git", branch: "main"),
.package(path: "../shared/ClawdbotKit"),
.package(path: "../../Swabble"),
.package(path: "../../Peekaboo/Core/PeekabooCore"),
.package(path: "../../Peekaboo/Core/PeekabooAutomationKit"),
],
targets: [
.target(
@ -61,8 +60,8 @@ let package = Package(
.product(name: "Subprocess", package: "swift-subprocess"),
.product(name: "Logging", package: "swift-log"),
.product(name: "Sparkle", package: "Sparkle"),
.product(name: "PeekabooBridge", package: "PeekabooCore"),
.product(name: "PeekabooAutomationKit", package: "PeekabooAutomationKit"),
.product(name: "PeekabooBridge", package: "Peekaboo"),
.product(name: "PeekabooAutomationKit", package: "Peekaboo"),
],
exclude: [
"Resources/Info.plist",

64
apps/macos/README.md Normal file
View File

@ -0,0 +1,64 @@
# Clawdbot macOS app (dev + signing)
## Quick dev run
```bash
# from repo root
scripts/restart-mac.sh
```
Options:
```bash
scripts/restart-mac.sh --no-sign # fastest dev; ad-hoc signing (TCC permissions do not stick)
scripts/restart-mac.sh --sign # force code signing (requires cert)
```
## Packaging flow
```bash
scripts/package-mac-app.sh
```
Creates `dist/Clawdbot.app` and signs it via `scripts/codesign-mac-app.sh`.
## Signing behavior
Auto-selects identity (first match):
1) Developer ID Application
2) Apple Distribution
3) Apple Development
4) first available identity
If none found:
- errors by default
- set `ALLOW_ADHOC_SIGNING=1` or `SIGN_IDENTITY="-"` to ad-hoc sign
## Team ID audit (Sparkle mismatch guard)
After signing, we read the app bundle Team ID and compare every Mach-O inside the app.
If any embedded binary has a different Team ID, signing fails.
Skip the audit:
```bash
SKIP_TEAM_ID_CHECK=1 scripts/package-mac-app.sh
```
## Library validation workaround (dev only)
If Sparkle Team ID mismatch blocks loading (common with Apple Development certs), opt in:
```bash
DISABLE_LIBRARY_VALIDATION=1 scripts/package-mac-app.sh
```
This adds `com.apple.security.cs.disable-library-validation` to app entitlements.
Use for local dev only; keep off for release builds.
## Useful env flags
- `SIGN_IDENTITY="Apple Development: Your Name (TEAMID)"`
- `ALLOW_ADHOC_SIGNING=1` (ad-hoc, TCC permissions do not persist)
- `CODESIGN_TIMESTAMP=off` (offline debug)
- `DISABLE_LIBRARY_VALIDATION=1` (dev-only Sparkle workaround)
- `SKIP_TEAM_ID_CHECK=1` (bypass audit)

View File

@ -170,8 +170,15 @@ final class AppState {
didSet { self.ifNotPreview { UserDefaults.standard.set(self.canvasEnabled, forKey: canvasEnabledKey) } }
}
var systemRunPolicy: SystemRunPolicy {
didSet { self.ifNotPreview { MacNodeConfigFile.setSystemRunPolicy(self.systemRunPolicy) } }
var execApprovalMode: ExecApprovalQuickMode {
didSet {
self.ifNotPreview {
ExecApprovalsStore.updateDefaults { defaults in
defaults.security = self.execApprovalMode.security
defaults.ask = self.execApprovalMode.ask
}
}
}
}
/// Tracks whether the Canvas panel is currently visible (not persisted).
@ -274,7 +281,8 @@ final class AppState {
self.remoteProjectRoot = UserDefaults.standard.string(forKey: remoteProjectRootKey) ?? ""
self.remoteCliPath = UserDefaults.standard.string(forKey: remoteCliPathKey) ?? ""
self.canvasEnabled = UserDefaults.standard.object(forKey: canvasEnabledKey) as? Bool ?? true
self.systemRunPolicy = SystemRunPolicy.load()
let execDefaults = ExecApprovalsStore.resolveDefaults()
self.execApprovalMode = ExecApprovalQuickMode.from(security: execDefaults.security, ask: execDefaults.ask)
self.peekabooBridgeEnabled = UserDefaults.standard
.object(forKey: peekabooBridgeEnabledKey) as? Bool ?? true
if !self.isPreview {

View File

@ -8,6 +8,8 @@ struct BridgeNodeInfo: Sendable {
var displayName: String?
var platform: String?
var version: String?
var coreVersion: String?
var uiVersion: String?
var deviceFamily: String?
var modelIdentifier: String?
var remoteAddress: String?
@ -147,6 +149,8 @@ actor BridgeConnectionHandler {
displayName: hello.displayName,
platform: hello.platform,
version: hello.version,
coreVersion: hello.coreVersion,
uiVersion: hello.uiVersion,
deviceFamily: hello.deviceFamily,
modelIdentifier: hello.modelIdentifier,
remoteAddress: self.remoteAddressString(),
@ -171,6 +175,8 @@ actor BridgeConnectionHandler {
displayName: req.displayName,
platform: req.platform,
version: req.version,
coreVersion: req.coreVersion,
uiVersion: req.uiVersion,
deviceFamily: req.deviceFamily,
modelIdentifier: req.modelIdentifier,
caps: req.caps,
@ -186,6 +192,8 @@ actor BridgeConnectionHandler {
displayName: enriched.displayName,
platform: enriched.platform,
version: enriched.version,
coreVersion: enriched.coreVersion,
uiVersion: enriched.uiVersion,
deviceFamily: enriched.deviceFamily,
modelIdentifier: enriched.modelIdentifier,
remoteAddress: enriched.remoteAddress,

View File

@ -214,9 +214,10 @@ enum CommandResolver {
subcommand: String,
extraArgs: [String] = [],
defaults: UserDefaults = .standard,
configRoot: [String: Any]? = nil,
searchPaths: [String]? = nil) -> [String]
{
let settings = self.connectionSettings(defaults: defaults)
let settings = self.connectionSettings(defaults: defaults, configRoot: configRoot)
if settings.mode == .remote, let ssh = self.sshNodeCommand(
subcommand: subcommand,
extraArgs: extraArgs,
@ -264,12 +265,14 @@ enum CommandResolver {
subcommand: String,
extraArgs: [String] = [],
defaults: UserDefaults = .standard,
configRoot: [String: Any]? = nil,
searchPaths: [String]? = nil) -> [String]
{
self.clawdbotNodeCommand(
subcommand: subcommand,
extraArgs: extraArgs,
defaults: defaults,
configRoot: configRoot,
searchPaths: searchPaths)
}
@ -384,8 +387,11 @@ enum CommandResolver {
let cliPath: String
}
static func connectionSettings(defaults: UserDefaults = .standard) -> RemoteSettings {
let root = ClawdbotConfigFile.loadDict()
static func connectionSettings(
defaults: UserDefaults = .standard,
configRoot: [String: Any]? = nil) -> RemoteSettings
{
let root = configRoot ?? ClawdbotConfigFile.loadDict()
let mode = ConnectionModeResolver.resolve(root: root, defaults: defaults).mode
let target = defaults.string(forKey: remoteTargetKey) ?? ""
let identity = defaults.string(forKey: remoteIdentityKey) ?? ""

View File

@ -0,0 +1,673 @@
import CryptoKit
import Foundation
import OSLog
import Security
enum ExecSecurity: String, CaseIterable, Codable, Identifiable {
case deny
case allowlist
case full
var id: String { self.rawValue }
var title: String {
switch self {
case .deny: "Deny"
case .allowlist: "Allowlist"
case .full: "Always Allow"
}
}
}
enum ExecApprovalQuickMode: String, CaseIterable, Identifiable {
case deny
case ask
case allow
var id: String { self.rawValue }
var title: String {
switch self {
case .deny: "Deny"
case .ask: "Always Ask"
case .allow: "Always Allow"
}
}
var security: ExecSecurity {
switch self {
case .deny: .deny
case .ask: .allowlist
case .allow: .full
}
}
var ask: ExecAsk {
switch self {
case .deny: .off
case .ask: .onMiss
case .allow: .off
}
}
static func from(security: ExecSecurity, ask: ExecAsk) -> ExecApprovalQuickMode {
switch security {
case .deny:
return .deny
case .full:
return .allow
case .allowlist:
return .ask
}
}
}
enum ExecAsk: String, CaseIterable, Codable, Identifiable {
case off
case onMiss = "on-miss"
case always
var id: String { self.rawValue }
var title: String {
switch self {
case .off: "Never Ask"
case .onMiss: "Ask on Allowlist Miss"
case .always: "Always Ask"
}
}
}
enum ExecApprovalDecision: String, Codable, Sendable {
case allowOnce = "allow-once"
case allowAlways = "allow-always"
case deny
}
struct ExecAllowlistEntry: Codable, Hashable {
var pattern: String
var lastUsedAt: Double? = nil
var lastUsedCommand: String? = nil
var lastResolvedPath: String? = nil
}
struct ExecApprovalsDefaults: Codable {
var security: ExecSecurity?
var ask: ExecAsk?
var askFallback: ExecSecurity?
var autoAllowSkills: Bool?
}
struct ExecApprovalsAgent: Codable {
var security: ExecSecurity?
var ask: ExecAsk?
var askFallback: ExecSecurity?
var autoAllowSkills: Bool?
var allowlist: [ExecAllowlistEntry]?
var isEmpty: Bool {
security == nil && ask == nil && askFallback == nil && autoAllowSkills == nil && (allowlist?.isEmpty ?? true)
}
}
struct ExecApprovalsSocketConfig: Codable {
var path: String?
var token: String?
}
struct ExecApprovalsFile: Codable {
var version: Int
var socket: ExecApprovalsSocketConfig?
var defaults: ExecApprovalsDefaults?
var agents: [String: ExecApprovalsAgent]?
}
struct ExecApprovalsSnapshot: Codable {
var path: String
var exists: Bool
var hash: String
var file: ExecApprovalsFile
}
struct ExecApprovalsResolved {
let url: URL
let socketPath: String
let token: String
let defaults: ExecApprovalsResolvedDefaults
let agent: ExecApprovalsResolvedDefaults
let allowlist: [ExecAllowlistEntry]
var file: ExecApprovalsFile
}
struct ExecApprovalsResolvedDefaults {
var security: ExecSecurity
var ask: ExecAsk
var askFallback: ExecSecurity
var autoAllowSkills: Bool
}
enum ExecApprovalsStore {
private static let logger = Logger(subsystem: "com.clawdbot", category: "exec-approvals")
private static let defaultSecurity: ExecSecurity = .deny
private static let defaultAsk: ExecAsk = .onMiss
private static let defaultAskFallback: ExecSecurity = .deny
private static let defaultAutoAllowSkills = false
static func fileURL() -> URL {
ClawdbotPaths.stateDirURL.appendingPathComponent("exec-approvals.json")
}
static func socketPath() -> String {
ClawdbotPaths.stateDirURL.appendingPathComponent("exec-approvals.sock").path
}
static func normalizeIncoming(_ file: ExecApprovalsFile) -> ExecApprovalsFile {
let socketPath = file.socket?.path?.trimmingCharacters(in: .whitespacesAndNewlines) ?? ""
let token = file.socket?.token?.trimmingCharacters(in: .whitespacesAndNewlines) ?? ""
return ExecApprovalsFile(
version: 1,
socket: ExecApprovalsSocketConfig(
path: socketPath.isEmpty ? nil : socketPath,
token: token.isEmpty ? nil : token),
defaults: file.defaults,
agents: file.agents)
}
static func readSnapshot() -> ExecApprovalsSnapshot {
let url = self.fileURL()
guard FileManager.default.fileExists(atPath: url.path) else {
return ExecApprovalsSnapshot(
path: url.path,
exists: false,
hash: self.hashRaw(nil),
file: ExecApprovalsFile(version: 1, socket: nil, defaults: nil, agents: [:]))
}
let raw = try? String(contentsOf: url, encoding: .utf8)
let data = raw.flatMap { $0.data(using: .utf8) }
let decoded: ExecApprovalsFile = {
if let data, let file = try? JSONDecoder().decode(ExecApprovalsFile.self, from: data), file.version == 1 {
return file
}
return ExecApprovalsFile(version: 1, socket: nil, defaults: nil, agents: [:])
}()
return ExecApprovalsSnapshot(
path: url.path,
exists: true,
hash: self.hashRaw(raw),
file: decoded)
}
static func redactForSnapshot(_ file: ExecApprovalsFile) -> ExecApprovalsFile {
let socketPath = file.socket?.path?.trimmingCharacters(in: .whitespacesAndNewlines) ?? ""
if socketPath.isEmpty {
return ExecApprovalsFile(
version: file.version,
socket: nil,
defaults: file.defaults,
agents: file.agents)
}
return ExecApprovalsFile(
version: file.version,
socket: ExecApprovalsSocketConfig(path: socketPath, token: nil),
defaults: file.defaults,
agents: file.agents)
}
static func loadFile() -> ExecApprovalsFile {
let url = self.fileURL()
guard FileManager.default.fileExists(atPath: url.path) else {
return ExecApprovalsFile(version: 1, socket: nil, defaults: nil, agents: [:])
}
do {
let data = try Data(contentsOf: url)
let decoded = try JSONDecoder().decode(ExecApprovalsFile.self, from: data)
if decoded.version != 1 {
return ExecApprovalsFile(version: 1, socket: nil, defaults: nil, agents: [:])
}
return decoded
} catch {
self.logger.warning("exec approvals load failed: \(error.localizedDescription, privacy: .public)")
return ExecApprovalsFile(version: 1, socket: nil, defaults: nil, agents: [:])
}
}
static func saveFile(_ file: ExecApprovalsFile) {
do {
let encoder = JSONEncoder()
encoder.outputFormatting = [.prettyPrinted, .sortedKeys]
let data = try encoder.encode(file)
let url = self.fileURL()
try FileManager.default.createDirectory(
at: url.deletingLastPathComponent(),
withIntermediateDirectories: true)
try data.write(to: url, options: [.atomic])
try? FileManager.default.setAttributes([.posixPermissions: 0o600], ofItemAtPath: url.path)
} catch {
self.logger.error("exec approvals save failed: \(error.localizedDescription, privacy: .public)")
}
}
static func ensureFile() -> ExecApprovalsFile {
var file = self.loadFile()
if file.socket == nil { file.socket = ExecApprovalsSocketConfig(path: nil, token: nil) }
let path = file.socket?.path?.trimmingCharacters(in: .whitespacesAndNewlines) ?? ""
if path.isEmpty {
file.socket?.path = self.socketPath()
}
let token = file.socket?.token?.trimmingCharacters(in: .whitespacesAndNewlines) ?? ""
if token.isEmpty {
file.socket?.token = self.generateToken()
}
if file.agents == nil { file.agents = [:] }
self.saveFile(file)
return file
}
static func resolve(agentId: String?) -> ExecApprovalsResolved {
let file = self.ensureFile()
let defaults = file.defaults ?? ExecApprovalsDefaults()
let resolvedDefaults = ExecApprovalsResolvedDefaults(
security: defaults.security ?? self.defaultSecurity,
ask: defaults.ask ?? self.defaultAsk,
askFallback: defaults.askFallback ?? self.defaultAskFallback,
autoAllowSkills: defaults.autoAllowSkills ?? self.defaultAutoAllowSkills)
let key = (agentId?.trimmingCharacters(in: .whitespacesAndNewlines).isEmpty == false)
? agentId!.trimmingCharacters(in: .whitespacesAndNewlines)
: "default"
let agentEntry = file.agents?[key] ?? ExecApprovalsAgent()
let resolvedAgent = ExecApprovalsResolvedDefaults(
security: agentEntry.security ?? resolvedDefaults.security,
ask: agentEntry.ask ?? resolvedDefaults.ask,
askFallback: agentEntry.askFallback ?? resolvedDefaults.askFallback,
autoAllowSkills: agentEntry.autoAllowSkills ?? resolvedDefaults.autoAllowSkills)
let allowlist = (agentEntry.allowlist ?? [])
.map { entry in
ExecAllowlistEntry(
pattern: entry.pattern.trimmingCharacters(in: .whitespacesAndNewlines),
lastUsedAt: entry.lastUsedAt,
lastUsedCommand: entry.lastUsedCommand,
lastResolvedPath: entry.lastResolvedPath)
}
.filter { !$0.pattern.isEmpty }
let socketPath = self.expandPath(file.socket?.path ?? self.socketPath())
let token = file.socket?.token ?? ""
return ExecApprovalsResolved(
url: self.fileURL(),
socketPath: socketPath,
token: token,
defaults: resolvedDefaults,
agent: resolvedAgent,
allowlist: allowlist,
file: file)
}
static func resolveDefaults() -> ExecApprovalsResolvedDefaults {
let file = self.ensureFile()
let defaults = file.defaults ?? ExecApprovalsDefaults()
return ExecApprovalsResolvedDefaults(
security: defaults.security ?? self.defaultSecurity,
ask: defaults.ask ?? self.defaultAsk,
askFallback: defaults.askFallback ?? self.defaultAskFallback,
autoAllowSkills: defaults.autoAllowSkills ?? self.defaultAutoAllowSkills)
}
static func saveDefaults(_ defaults: ExecApprovalsDefaults) {
self.updateFile { file in
file.defaults = defaults
}
}
static func updateDefaults(_ mutate: (inout ExecApprovalsDefaults) -> Void) {
self.updateFile { file in
var defaults = file.defaults ?? ExecApprovalsDefaults()
mutate(&defaults)
file.defaults = defaults
}
}
static func saveAgent(_ agent: ExecApprovalsAgent, agentId: String?) {
self.updateFile { file in
var agents = file.agents ?? [:]
let key = self.agentKey(agentId)
if agent.isEmpty {
agents.removeValue(forKey: key)
} else {
agents[key] = agent
}
file.agents = agents.isEmpty ? nil : agents
}
}
static func addAllowlistEntry(agentId: String?, pattern: String) {
let trimmed = pattern.trimmingCharacters(in: .whitespacesAndNewlines)
guard !trimmed.isEmpty else { return }
self.updateFile { file in
let key = self.agentKey(agentId)
var agents = file.agents ?? [:]
var entry = agents[key] ?? ExecApprovalsAgent()
var allowlist = entry.allowlist ?? []
if allowlist.contains(where: { $0.pattern == trimmed }) { return }
allowlist.append(ExecAllowlistEntry(pattern: trimmed, lastUsedAt: Date().timeIntervalSince1970 * 1000))
entry.allowlist = allowlist
agents[key] = entry
file.agents = agents
}
}
static func recordAllowlistUse(
agentId: String?,
pattern: String,
command: String,
resolvedPath: String?)
{
self.updateFile { file in
let key = self.agentKey(agentId)
var agents = file.agents ?? [:]
var entry = agents[key] ?? ExecApprovalsAgent()
let allowlist = (entry.allowlist ?? []).map { item -> ExecAllowlistEntry in
guard item.pattern == pattern else { return item }
return ExecAllowlistEntry(
pattern: item.pattern,
lastUsedAt: Date().timeIntervalSince1970 * 1000,
lastUsedCommand: command,
lastResolvedPath: resolvedPath)
}
entry.allowlist = allowlist
agents[key] = entry
file.agents = agents
}
}
static func updateAllowlist(agentId: String?, allowlist: [ExecAllowlistEntry]) {
self.updateFile { file in
let key = self.agentKey(agentId)
var agents = file.agents ?? [:]
var entry = agents[key] ?? ExecApprovalsAgent()
let cleaned = allowlist
.map { item in
ExecAllowlistEntry(
pattern: item.pattern.trimmingCharacters(in: .whitespacesAndNewlines),
lastUsedAt: item.lastUsedAt,
lastUsedCommand: item.lastUsedCommand,
lastResolvedPath: item.lastResolvedPath)
}
.filter { !$0.pattern.isEmpty }
entry.allowlist = cleaned
agents[key] = entry
file.agents = agents
}
}
static func updateAgentSettings(agentId: String?, mutate: (inout ExecApprovalsAgent) -> Void) {
self.updateFile { file in
let key = self.agentKey(agentId)
var agents = file.agents ?? [:]
var entry = agents[key] ?? ExecApprovalsAgent()
mutate(&entry)
if entry.isEmpty {
agents.removeValue(forKey: key)
} else {
agents[key] = entry
}
file.agents = agents.isEmpty ? nil : agents
}
}
private static func updateFile(_ mutate: (inout ExecApprovalsFile) -> Void) {
var file = self.ensureFile()
mutate(&file)
self.saveFile(file)
}
private static func generateToken() -> String {
var bytes = [UInt8](repeating: 0, count: 24)
let status = SecRandomCopyBytes(kSecRandomDefault, bytes.count, &bytes)
if status == errSecSuccess {
return Data(bytes)
.base64EncodedString()
.replacingOccurrences(of: "+", with: "-")
.replacingOccurrences(of: "/", with: "_")
.replacingOccurrences(of: "=", with: "")
}
return UUID().uuidString
}
private static func hashRaw(_ raw: String?) -> String {
let data = Data((raw ?? "").utf8)
let digest = SHA256.hash(data: data)
return digest.map { String(format: "%02x", $0) }.joined()
}
private static func expandPath(_ raw: String) -> String {
let trimmed = raw.trimmingCharacters(in: .whitespacesAndNewlines)
if trimmed == "~" {
return FileManager.default.homeDirectoryForCurrentUser.path
}
if trimmed.hasPrefix("~/") {
let suffix = trimmed.dropFirst(2)
return FileManager.default.homeDirectoryForCurrentUser
.appendingPathComponent(String(suffix)).path
}
return trimmed
}
private static func agentKey(_ agentId: String?) -> String {
let trimmed = agentId?.trimmingCharacters(in: .whitespacesAndNewlines) ?? ""
return trimmed.isEmpty ? "default" : trimmed
}
}
struct ExecCommandResolution: Sendable {
let rawExecutable: String
let resolvedPath: String?
let executableName: String
let cwd: String?
static func resolve(
command: [String],
rawCommand: String?,
cwd: String?,
env: [String: String]?
) -> ExecCommandResolution? {
let trimmedRaw = rawCommand?.trimmingCharacters(in: .whitespacesAndNewlines) ?? ""
if !trimmedRaw.isEmpty, let token = self.parseFirstToken(trimmedRaw) {
return self.resolveExecutable(rawExecutable: token, cwd: cwd, env: env)
}
return self.resolve(command: command, cwd: cwd, env: env)
}
static func resolve(command: [String], cwd: String?, env: [String: String]?) -> ExecCommandResolution? {
guard let raw = command.first?.trimmingCharacters(in: .whitespacesAndNewlines), !raw.isEmpty else {
return nil
}
return self.resolveExecutable(rawExecutable: raw, cwd: cwd, env: env)
}
private static func resolveExecutable(
rawExecutable: String,
cwd: String?,
env: [String: String]?
) -> ExecCommandResolution? {
let expanded = rawExecutable.hasPrefix("~") ? (rawExecutable as NSString).expandingTildeInPath : rawExecutable
let hasPathSeparator = expanded.contains("/") || expanded.contains("\\")
let resolvedPath: String? = {
if hasPathSeparator {
if expanded.hasPrefix("/") {
return expanded
}
let base = cwd?.trimmingCharacters(in: .whitespacesAndNewlines)
let root = (base?.isEmpty == false) ? base! : FileManager.default.currentDirectoryPath
return URL(fileURLWithPath: root).appendingPathComponent(expanded).path
}
let searchPaths = self.searchPaths(from: env)
return CommandResolver.findExecutable(named: expanded, searchPaths: searchPaths)
}()
let name = resolvedPath.map { URL(fileURLWithPath: $0).lastPathComponent } ?? expanded
return ExecCommandResolution(rawExecutable: expanded, resolvedPath: resolvedPath, executableName: name, cwd: cwd)
}
private static func parseFirstToken(_ command: String) -> String? {
let trimmed = command.trimmingCharacters(in: .whitespacesAndNewlines)
guard !trimmed.isEmpty else { return nil }
guard let first = trimmed.first else { return nil }
if first == "\"" || first == "'" {
let rest = trimmed.dropFirst()
if let end = rest.firstIndex(of: first) {
return String(rest[..<end])
}
return String(rest)
}
return trimmed.split(whereSeparator: { $0.isWhitespace }).first.map(String.init)
}
private static func searchPaths(from env: [String: String]?) -> [String] {
let raw = env?["PATH"]
if let raw, !raw.isEmpty {
return raw.split(separator: ":").map(String.init)
}
return CommandResolver.preferredPaths()
}
}
enum ExecCommandFormatter {
static func displayString(for argv: [String]) -> String {
argv.map { arg in
let trimmed = arg.trimmingCharacters(in: .whitespacesAndNewlines)
guard !trimmed.isEmpty else { return "\"\"" }
let needsQuotes = trimmed.contains { $0.isWhitespace || $0 == "\"" }
if !needsQuotes { return trimmed }
let escaped = trimmed.replacingOccurrences(of: "\"", with: "\\\"")
return "\"\(escaped)\""
}.joined(separator: " ")
}
static func displayString(for argv: [String], rawCommand: String?) -> String {
let trimmed = rawCommand?.trimmingCharacters(in: .whitespacesAndNewlines) ?? ""
if !trimmed.isEmpty { return trimmed }
return self.displayString(for: argv)
}
}
enum ExecAllowlistMatcher {
static func match(entries: [ExecAllowlistEntry], resolution: ExecCommandResolution?) -> ExecAllowlistEntry? {
guard let resolution, !entries.isEmpty else { return nil }
let rawExecutable = resolution.rawExecutable
let resolvedPath = resolution.resolvedPath
let executableName = resolution.executableName
for entry in entries {
let pattern = entry.pattern.trimmingCharacters(in: .whitespacesAndNewlines)
if pattern.isEmpty { continue }
let hasPath = pattern.contains("/") || pattern.contains("~") || pattern.contains("\\")
if hasPath {
let target = resolvedPath ?? rawExecutable
if self.matches(pattern: pattern, target: target) { return entry }
} else if self.matches(pattern: pattern, target: executableName) {
return entry
}
}
return nil
}
private static func matches(pattern: String, target: String) -> Bool {
let trimmed = pattern.trimmingCharacters(in: .whitespacesAndNewlines)
guard !trimmed.isEmpty else { return false }
let expanded = trimmed.hasPrefix("~") ? (trimmed as NSString).expandingTildeInPath : trimmed
let normalizedPattern = self.normalizeMatchTarget(expanded)
let normalizedTarget = self.normalizeMatchTarget(target)
guard let regex = self.regex(for: normalizedPattern) else { return false }
let range = NSRange(location: 0, length: normalizedTarget.utf16.count)
return regex.firstMatch(in: normalizedTarget, options: [], range: range) != nil
}
private static func normalizeMatchTarget(_ value: String) -> String {
value.replacingOccurrences(of: "\\\\", with: "/").lowercased()
}
private static func regex(for pattern: String) -> NSRegularExpression? {
var regex = "^"
var idx = pattern.startIndex
while idx < pattern.endIndex {
let ch = pattern[idx]
if ch == "*" {
let next = pattern.index(after: idx)
if next < pattern.endIndex, pattern[next] == "*" {
regex += ".*"
idx = pattern.index(after: next)
} else {
regex += "[^/]*"
idx = next
}
continue
}
if ch == "?" {
regex += "."
idx = pattern.index(after: idx)
continue
}
regex += NSRegularExpression.escapedPattern(for: String(ch))
idx = pattern.index(after: idx)
}
regex += "$"
return try? NSRegularExpression(pattern: regex, options: [.caseInsensitive])
}
}
struct ExecEventPayload: Codable, Sendable {
var sessionKey: String
var runId: String
var host: String
var command: String?
var exitCode: Int?
var timedOut: Bool?
var success: Bool?
var output: String?
var reason: String?
static func truncateOutput(_ raw: String, maxChars: Int = 20_000) -> String? {
let trimmed = raw.trimmingCharacters(in: .whitespacesAndNewlines)
guard !trimmed.isEmpty else { return nil }
if trimmed.count <= maxChars { return trimmed }
let suffix = trimmed.suffix(maxChars)
return "... (truncated) \(suffix)"
}
}
actor SkillBinsCache {
static let shared = SkillBinsCache()
private var bins: Set<String> = []
private var lastRefresh: Date?
private let refreshInterval: TimeInterval = 90
func currentBins(force: Bool = false) async -> Set<String> {
if force || self.isStale() {
await self.refresh()
}
return self.bins
}
func refresh() async {
do {
let report = try await GatewayConnection.shared.skillsStatus()
var next = Set<String>()
for skill in report.skills {
for bin in skill.requirements.bins {
let trimmed = bin.trimmingCharacters(in: .whitespacesAndNewlines)
if !trimmed.isEmpty { next.insert(trimmed) }
}
}
self.bins = next
self.lastRefresh = Date()
} catch {
if self.lastRefresh == nil {
self.bins = []
}
}
}
private func isStale() -> Bool {
guard let lastRefresh else { return true }
return Date().timeIntervalSince(lastRefresh) > self.refreshInterval
}
}

View File

@ -0,0 +1,666 @@
import AppKit
import ClawdbotKit
import CryptoKit
import Darwin
import Foundation
import OSLog
struct ExecApprovalPromptRequest: Codable, Sendable {
var command: String
var cwd: String?
var host: String?
var security: String?
var ask: String?
var agentId: String?
var resolvedPath: String?
}
private struct ExecApprovalSocketRequest: Codable {
var type: String
var token: String
var id: String
var request: ExecApprovalPromptRequest
}
private struct ExecApprovalSocketDecision: Codable {
var type: String
var id: String
var decision: ExecApprovalDecision
}
fileprivate struct ExecHostSocketRequest: Codable {
var type: String
var id: String
var nonce: String
var ts: Int
var hmac: String
var requestJson: String
}
fileprivate struct ExecHostRequest: Codable {
var command: [String]
var rawCommand: String?
var cwd: String?
var env: [String: String]?
var timeoutMs: Int?
var needsScreenRecording: Bool?
var agentId: String?
var sessionKey: String?
}
fileprivate struct ExecHostRunResult: Codable {
var exitCode: Int?
var timedOut: Bool
var success: Bool
var stdout: String
var stderr: String
var error: String?
}
fileprivate struct ExecHostError: Codable {
var code: String
var message: String
var reason: String?
}
fileprivate struct ExecHostResponse: Codable {
var type: String
var id: String
var ok: Bool
var payload: ExecHostRunResult?
var error: ExecHostError?
}
enum ExecApprovalsSocketClient {
private struct TimeoutError: LocalizedError {
var message: String
var errorDescription: String? { message }
}
static func requestDecision(
socketPath: String,
token: String,
request: ExecApprovalPromptRequest,
timeoutMs: Int = 15_000) async -> ExecApprovalDecision?
{
let trimmedPath = socketPath.trimmingCharacters(in: .whitespacesAndNewlines)
let trimmedToken = token.trimmingCharacters(in: .whitespacesAndNewlines)
guard !trimmedPath.isEmpty, !trimmedToken.isEmpty else { return nil }
do {
return try await AsyncTimeout.withTimeoutMs(timeoutMs: timeoutMs, onTimeout: {
TimeoutError(message: "exec approvals socket timeout")
}, operation: {
try await Task.detached {
try self.requestDecisionSync(
socketPath: trimmedPath,
token: trimmedToken,
request: request)
}.value
})
} catch {
return nil
}
}
private static func requestDecisionSync(
socketPath: String,
token: String,
request: ExecApprovalPromptRequest) throws -> ExecApprovalDecision?
{
let fd = socket(AF_UNIX, SOCK_STREAM, 0)
guard fd >= 0 else {
throw NSError(domain: "ExecApprovals", code: 1, userInfo: [
NSLocalizedDescriptionKey: "socket create failed",
])
}
var addr = sockaddr_un()
addr.sun_family = sa_family_t(AF_UNIX)
let maxLen = MemoryLayout.size(ofValue: addr.sun_path)
if socketPath.utf8.count >= maxLen {
throw NSError(domain: "ExecApprovals", code: 2, userInfo: [
NSLocalizedDescriptionKey: "socket path too long",
])
}
socketPath.withCString { cstr in
withUnsafeMutablePointer(to: &addr.sun_path) { ptr in
let raw = UnsafeMutableRawPointer(ptr).assumingMemoryBound(to: Int8.self)
strncpy(raw, cstr, maxLen - 1)
}
}
let size = socklen_t(MemoryLayout.size(ofValue: addr))
let result = withUnsafePointer(to: &addr) { ptr in
ptr.withMemoryRebound(to: sockaddr.self, capacity: 1) { rebound in
connect(fd, rebound, size)
}
}
if result != 0 {
throw NSError(domain: "ExecApprovals", code: 3, userInfo: [
NSLocalizedDescriptionKey: "socket connect failed",
])
}
let handle = FileHandle(fileDescriptor: fd, closeOnDealloc: true)
let message = ExecApprovalSocketRequest(
type: "request",
token: token,
id: UUID().uuidString,
request: request)
let data = try JSONEncoder().encode(message)
var payload = data
payload.append(0x0A)
try handle.write(contentsOf: payload)
guard let line = try self.readLine(from: handle, maxBytes: 256_000),
let lineData = line.data(using: .utf8)
else { return nil }
let response = try JSONDecoder().decode(ExecApprovalSocketDecision.self, from: lineData)
return response.decision
}
private static func readLine(from handle: FileHandle, maxBytes: Int) throws -> String? {
var buffer = Data()
while buffer.count < maxBytes {
let chunk = try handle.read(upToCount: 4096) ?? Data()
if chunk.isEmpty { break }
buffer.append(chunk)
if buffer.contains(0x0A) { break }
}
guard let newlineIndex = buffer.firstIndex(of: 0x0A) else {
guard !buffer.isEmpty else { return nil }
return String(data: buffer, encoding: .utf8)
}
let lineData = buffer.subdata(in: 0..<newlineIndex)
return String(data: lineData, encoding: .utf8)
}
}
@MainActor
final class ExecApprovalsPromptServer {
static let shared = ExecApprovalsPromptServer()
private var server: ExecApprovalsSocketServer?
func start() {
guard self.server == nil else { return }
let approvals = ExecApprovalsStore.resolve(agentId: nil)
let server = ExecApprovalsSocketServer(
socketPath: approvals.socketPath,
token: approvals.token,
onPrompt: { request in
await ExecApprovalsPromptPresenter.prompt(request)
},
onExec: { request in
await ExecHostExecutor.handle(request)
})
server.start()
self.server = server
}
func stop() {
self.server?.stop()
self.server = nil
}
}
enum ExecApprovalsPromptPresenter {
@MainActor
static func prompt(_ request: ExecApprovalPromptRequest) -> ExecApprovalDecision {
NSApp.activate(ignoringOtherApps: true)
let alert = NSAlert()
alert.alertStyle = .warning
alert.messageText = "Allow this command?"
var details = "Clawdbot wants to run:\n\n\(request.command)"
let trimmedCwd = request.cwd?.trimmingCharacters(in: .whitespacesAndNewlines) ?? ""
if !trimmedCwd.isEmpty {
details += "\n\nWorking directory:\n\(trimmedCwd)"
}
let trimmedAgent = request.agentId?.trimmingCharacters(in: .whitespacesAndNewlines) ?? ""
if !trimmedAgent.isEmpty {
details += "\n\nAgent:\n\(trimmedAgent)"
}
let trimmedPath = request.resolvedPath?.trimmingCharacters(in: .whitespacesAndNewlines) ?? ""
if !trimmedPath.isEmpty {
details += "\n\nExecutable:\n\(trimmedPath)"
}
let trimmedHost = request.host?.trimmingCharacters(in: .whitespacesAndNewlines) ?? ""
if !trimmedHost.isEmpty {
details += "\n\nHost:\n\(trimmedHost)"
}
if let security = request.security?.trimmingCharacters(in: .whitespacesAndNewlines), !security.isEmpty {
details += "\n\nSecurity:\n\(security)"
}
if let ask = request.ask?.trimmingCharacters(in: .whitespacesAndNewlines), !ask.isEmpty {
details += "\nAsk mode:\n\(ask)"
}
details += "\n\nThis runs on this machine."
alert.informativeText = details
alert.addButton(withTitle: "Allow Once")
alert.addButton(withTitle: "Always Allow")
alert.addButton(withTitle: "Don't Allow")
switch alert.runModal() {
case .alertFirstButtonReturn:
return .allowOnce
case .alertSecondButtonReturn:
return .allowAlways
default:
return .deny
}
}
}
@MainActor
fileprivate enum ExecHostExecutor {
private static let blockedEnvKeys: Set<String> = [
"PATH",
"NODE_OPTIONS",
"PYTHONHOME",
"PYTHONPATH",
"PERL5LIB",
"PERL5OPT",
"RUBYOPT",
]
private static let blockedEnvPrefixes: [String] = [
"DYLD_",
"LD_",
]
static func handle(_ request: ExecHostRequest) async -> ExecHostResponse {
let command = request.command.map { $0.trimmingCharacters(in: .whitespacesAndNewlines) }
guard !command.isEmpty else {
return ExecHostResponse(
type: "exec-res",
id: UUID().uuidString,
ok: false,
payload: nil,
error: ExecHostError(code: "INVALID_REQUEST", message: "command required", reason: "invalid"))
}
let displayCommand = ExecCommandFormatter.displayString(
for: command,
rawCommand: request.rawCommand)
let agentId = request.agentId?.trimmingCharacters(in: .whitespacesAndNewlines)
let trimmedAgent = (agentId?.isEmpty == false) ? agentId : nil
let approvals = ExecApprovalsStore.resolve(agentId: trimmedAgent)
let security = approvals.agent.security
let ask = approvals.agent.ask
let autoAllowSkills = approvals.agent.autoAllowSkills
let env = self.sanitizedEnv(request.env)
let resolution = ExecCommandResolution.resolve(
command: command,
rawCommand: request.rawCommand,
cwd: request.cwd,
env: env)
let allowlistMatch = security == .allowlist
? ExecAllowlistMatcher.match(entries: approvals.allowlist, resolution: resolution)
: nil
let skillAllow: Bool
if autoAllowSkills, let name = resolution?.executableName {
let bins = await SkillBinsCache.shared.currentBins()
skillAllow = bins.contains(name)
} else {
skillAllow = false
}
if security == .deny {
return ExecHostResponse(
type: "exec-res",
id: UUID().uuidString,
ok: false,
payload: nil,
error: ExecHostError(code: "UNAVAILABLE", message: "SYSTEM_RUN_DISABLED: security=deny", reason: "security=deny"))
}
let requiresAsk: Bool = {
if ask == .always { return true }
if ask == .onMiss && security == .allowlist && allowlistMatch == nil && !skillAllow { return true }
return false
}()
var approvedByAsk = false
if requiresAsk {
let decision = ExecApprovalsPromptPresenter.prompt(
ExecApprovalPromptRequest(
command: displayCommand,
cwd: request.cwd,
host: "node",
security: security.rawValue,
ask: ask.rawValue,
agentId: trimmedAgent,
resolvedPath: resolution?.resolvedPath))
switch decision {
case .deny:
return ExecHostResponse(
type: "exec-res",
id: UUID().uuidString,
ok: false,
payload: nil,
error: ExecHostError(code: "UNAVAILABLE", message: "SYSTEM_RUN_DENIED: user denied", reason: "user-denied"))
case .allowAlways:
approvedByAsk = true
if security == .allowlist {
let pattern = resolution?.resolvedPath ?? resolution?.rawExecutable ?? command.first ?? ""
if !pattern.isEmpty {
ExecApprovalsStore.addAllowlistEntry(agentId: trimmedAgent, pattern: pattern)
}
}
case .allowOnce:
approvedByAsk = true
}
}
if security == .allowlist && allowlistMatch == nil && !skillAllow && !approvedByAsk {
return ExecHostResponse(
type: "exec-res",
id: UUID().uuidString,
ok: false,
payload: nil,
error: ExecHostError(code: "UNAVAILABLE", message: "SYSTEM_RUN_DENIED: allowlist miss", reason: "allowlist-miss"))
}
if let match = allowlistMatch {
ExecApprovalsStore.recordAllowlistUse(
agentId: trimmedAgent,
pattern: match.pattern,
command: displayCommand,
resolvedPath: resolution?.resolvedPath)
}
if request.needsScreenRecording == true {
let authorized = await PermissionManager
.status([.screenRecording])[.screenRecording] ?? false
if !authorized {
return ExecHostResponse(
type: "exec-res",
id: UUID().uuidString,
ok: false,
payload: nil,
error: ExecHostError(code: "UNAVAILABLE", message: "PERMISSION_MISSING: screenRecording", reason: "permission:screenRecording"))
}
}
let timeoutSec = request.timeoutMs.flatMap { Double($0) / 1000.0 }
let result = await Task.detached { () -> ShellExecutor.ShellResult in
await ShellExecutor.runDetailed(
command: command,
cwd: request.cwd,
env: env,
timeout: timeoutSec)
}.value
let payload = ExecHostRunResult(
exitCode: result.exitCode,
timedOut: result.timedOut,
success: result.success,
stdout: result.stdout,
stderr: result.stderr,
error: result.errorMessage)
return ExecHostResponse(
type: "exec-res",
id: UUID().uuidString,
ok: true,
payload: payload,
error: nil)
}
private static func sanitizedEnv(_ overrides: [String: String]?) -> [String: String]? {
guard let overrides else { return nil }
var merged = ProcessInfo.processInfo.environment
for (rawKey, value) in overrides {
let key = rawKey.trimmingCharacters(in: .whitespacesAndNewlines)
guard !key.isEmpty else { continue }
let upper = key.uppercased()
if self.blockedEnvKeys.contains(upper) { continue }
if self.blockedEnvPrefixes.contains(where: { upper.hasPrefix($0) }) { continue }
merged[key] = value
}
return merged
}
}
private final class ExecApprovalsSocketServer: @unchecked Sendable {
private let logger = Logger(subsystem: "com.clawdbot", category: "exec-approvals.socket")
private let socketPath: String
private let token: String
private let onPrompt: @Sendable (ExecApprovalPromptRequest) async -> ExecApprovalDecision
private let onExec: @Sendable (ExecHostRequest) async -> ExecHostResponse
private var socketFD: Int32 = -1
private var acceptTask: Task<Void, Never>?
private var isRunning = false
init(
socketPath: String,
token: String,
onPrompt: @escaping @Sendable (ExecApprovalPromptRequest) async -> ExecApprovalDecision,
onExec: @escaping @Sendable (ExecHostRequest) async -> ExecHostResponse)
{
self.socketPath = socketPath
self.token = token
self.onPrompt = onPrompt
self.onExec = onExec
}
func start() {
guard !self.isRunning else { return }
self.isRunning = true
self.acceptTask = Task.detached { [weak self] in
await self?.runAcceptLoop()
}
}
func stop() {
self.isRunning = false
self.acceptTask?.cancel()
self.acceptTask = nil
if self.socketFD >= 0 {
close(self.socketFD)
self.socketFD = -1
}
if !self.socketPath.isEmpty {
unlink(self.socketPath)
}
}
private func runAcceptLoop() async {
let fd = self.openSocket()
guard fd >= 0 else {
self.isRunning = false
return
}
self.socketFD = fd
while self.isRunning {
var addr = sockaddr_un()
var len = socklen_t(MemoryLayout.size(ofValue: addr))
let client = withUnsafeMutablePointer(to: &addr) { ptr in
ptr.withMemoryRebound(to: sockaddr.self, capacity: 1) { rebound in
accept(fd, rebound, &len)
}
}
if client < 0 {
if errno == EINTR { continue }
break
}
Task.detached { [weak self] in
await self?.handleClient(fd: client)
}
}
}
private func openSocket() -> Int32 {
let fd = socket(AF_UNIX, SOCK_STREAM, 0)
guard fd >= 0 else {
self.logger.error("exec approvals socket create failed")
return -1
}
unlink(self.socketPath)
var addr = sockaddr_un()
addr.sun_family = sa_family_t(AF_UNIX)
let maxLen = MemoryLayout.size(ofValue: addr.sun_path)
if self.socketPath.utf8.count >= maxLen {
self.logger.error("exec approvals socket path too long")
close(fd)
return -1
}
self.socketPath.withCString { cstr in
withUnsafeMutablePointer(to: &addr.sun_path) { ptr in
let raw = UnsafeMutableRawPointer(ptr).assumingMemoryBound(to: Int8.self)
memset(raw, 0, maxLen)
strncpy(raw, cstr, maxLen - 1)
}
}
let size = socklen_t(MemoryLayout.size(ofValue: addr))
let result = withUnsafePointer(to: &addr) { ptr in
ptr.withMemoryRebound(to: sockaddr.self, capacity: 1) { rebound in
bind(fd, rebound, size)
}
}
if result != 0 {
self.logger.error("exec approvals socket bind failed")
close(fd)
return -1
}
if listen(fd, 16) != 0 {
self.logger.error("exec approvals socket listen failed")
close(fd)
return -1
}
chmod(self.socketPath, 0o600)
self.logger.info("exec approvals socket listening at \(self.socketPath, privacy: .public)")
return fd
}
private func handleClient(fd: Int32) async {
let handle = FileHandle(fileDescriptor: fd, closeOnDealloc: true)
do {
guard self.isAllowedPeer(fd: fd) else {
try self.sendApprovalResponse(handle: handle, id: UUID().uuidString, decision: .deny)
return
}
guard let line = try self.readLine(from: handle, maxBytes: 256_000),
let data = line.data(using: .utf8)
else {
return
}
guard
let envelope = try JSONSerialization.jsonObject(with: data) as? [String: Any],
let type = envelope["type"] as? String
else {
return
}
if type == "request" {
let request = try JSONDecoder().decode(ExecApprovalSocketRequest.self, from: data)
guard request.token == self.token else {
try self.sendApprovalResponse(handle: handle, id: request.id, decision: .deny)
return
}
let decision = await self.onPrompt(request.request)
try self.sendApprovalResponse(handle: handle, id: request.id, decision: decision)
return
}
if type == "exec" {
let request = try JSONDecoder().decode(ExecHostSocketRequest.self, from: data)
let response = await self.handleExecRequest(request)
try self.sendExecResponse(handle: handle, response: response)
return
}
} catch {
self.logger.error("exec approvals socket handling failed: \(error.localizedDescription, privacy: .public)")
}
}
private func readLine(from handle: FileHandle, maxBytes: Int) throws -> String? {
var buffer = Data()
while buffer.count < maxBytes {
let chunk = try handle.read(upToCount: 4096) ?? Data()
if chunk.isEmpty { break }
buffer.append(chunk)
if buffer.contains(0x0A) { break }
}
guard let newlineIndex = buffer.firstIndex(of: 0x0A) else {
guard !buffer.isEmpty else { return nil }
return String(data: buffer, encoding: .utf8)
}
let lineData = buffer.subdata(in: 0..<newlineIndex)
return String(data: lineData, encoding: .utf8)
}
private func sendApprovalResponse(
handle: FileHandle,
id: String,
decision: ExecApprovalDecision) throws
{
let response = ExecApprovalSocketDecision(type: "decision", id: id, decision: decision)
let data = try JSONEncoder().encode(response)
var payload = data
payload.append(0x0A)
try handle.write(contentsOf: payload)
}
private func sendExecResponse(handle: FileHandle, response: ExecHostResponse) throws {
let data = try JSONEncoder().encode(response)
var payload = data
payload.append(0x0A)
try handle.write(contentsOf: payload)
}
private func isAllowedPeer(fd: Int32) -> Bool {
var uid = uid_t(0)
var gid = gid_t(0)
if getpeereid(fd, &uid, &gid) != 0 {
return false
}
return uid == geteuid()
}
private func handleExecRequest(_ request: ExecHostSocketRequest) async -> ExecHostResponse {
let nowMs = Int(Date().timeIntervalSince1970 * 1000)
if abs(nowMs - request.ts) > 10_000 {
return ExecHostResponse(
type: "exec-res",
id: request.id,
ok: false,
payload: nil,
error: ExecHostError(code: "INVALID_REQUEST", message: "expired request", reason: "ttl"))
}
let expected = self.hmacHex(nonce: request.nonce, ts: request.ts, requestJson: request.requestJson)
if expected != request.hmac {
return ExecHostResponse(
type: "exec-res",
id: request.id,
ok: false,
payload: nil,
error: ExecHostError(code: "INVALID_REQUEST", message: "invalid auth", reason: "hmac"))
}
guard let requestData = request.requestJson.data(using: .utf8),
let payload = try? JSONDecoder().decode(ExecHostRequest.self, from: requestData)
else {
return ExecHostResponse(
type: "exec-res",
id: request.id,
ok: false,
payload: nil,
error: ExecHostError(code: "INVALID_REQUEST", message: "invalid payload", reason: "json"))
}
let response = await self.onExec(payload)
return ExecHostResponse(
type: "exec-res",
id: request.id,
ok: response.ok,
payload: response.payload,
error: response.error)
}
private func hmacHex(nonce: String, ts: Int, requestJson: String) -> String {
let key = SymmetricKey(data: Data(self.token.utf8))
let message = "\(nonce):\(ts):\(requestJson)"
let mac = HMAC<SHA256>.authenticationCode(for: Data(message.utf8), using: key)
return mac.map { String(format: "%02x", $0) }.joined()
}
}

View File

@ -249,6 +249,13 @@ actor GatewayConnection {
return trimmed.isEmpty ? nil : trimmed
}
func cachedGatewayVersion() -> String? {
guard let snapshot = self.lastSnapshot else { return nil }
let raw = snapshot.server["version"]?.value as? String
let trimmed = raw?.trimmingCharacters(in: CharacterSet.whitespacesAndNewlines) ?? ""
return trimmed.isEmpty ? nil : trimmed
}
func snapshotPaths() -> (configPath: String?, stateDir: String?) {
guard let snapshot = self.lastSnapshot else { return (nil, nil) }
let configPath = snapshot.snapshot.configpath?.trimmingCharacters(in: .whitespacesAndNewlines)

View File

@ -27,7 +27,11 @@ struct Semver: Comparable, CustomStringConvertible, Sendable {
else { return nil }
// Strip prerelease suffix (e.g., "11-4" "11", "5-beta.1" "5")
let patchRaw = String(parts[2])
let patchNumeric = patchRaw.split { $0 == "-" || $0 == "+" }.first.flatMap { Int($0) } ?? 0
guard let patchToken = patchRaw.split(whereSeparator: { $0 == "-" || $0 == "+" }).first,
let patchNumeric = Int(patchToken)
else {
return nil
}
return Semver(major: major, minor: minor, patch: patchNumeric)
}

View File

@ -16,6 +16,10 @@ enum GatewayLaunchAgentManager {
static func set(enabled: Bool, bundlePath: String, port: Int) async -> String? {
_ = bundlePath
guard !CommandResolver.connectionModeIsRemote() else {
self.logger.info("launchd change skipped (remote mode)")
return nil
}
if enabled, self.isLaunchAgentWriteDisabled() {
self.logger.info("launchd enable skipped (disable marker set)")
return nil
@ -112,7 +116,9 @@ extension GatewayLaunchAgentManager {
{
let command = CommandResolver.clawdbotCommand(
subcommand: "daemon",
extraArgs: self.withJsonFlag(args))
extraArgs: self.withJsonFlag(args),
// Launchd management must always run locally, even if remote mode is configured.
configRoot: ["gateway": ["mode": "local"]])
var env = ProcessInfo.processInfo.environment
env["PATH"] = CommandResolver.preferredPaths().joined(separator: ":")
let response = await ShellExecutor.runDetailed(command: command, cwd: nil, env: env, timeout: timeout)

View File

@ -114,6 +114,9 @@ final class GatewayProcessManager {
self.lastFailureReason = nil
self.status = .stopped
self.logger.info("gateway stop requested")
if CommandResolver.connectionModeIsRemote() {
return
}
let bundlePath = Bundle.main.bundleURL.path
Task {
_ = await GatewayLaunchAgentManager.set(

View File

@ -83,27 +83,7 @@ struct GeneralSettings: View {
subtitle: "Allow the agent to capture a photo or short video via the built-in camera.",
binding: self.$cameraEnabled)
VStack(alignment: .leading, spacing: 6) {
Text("Node Run Commands")
.font(.body)
Picker("", selection: self.$state.systemRunPolicy) {
ForEach(SystemRunPolicy.allCases) { policy in
Text(policy.title).tag(policy)
}
}
.labelsHidden()
.pickerStyle(.menu)
Text("""
Controls remote command execution on this Mac when it is paired as a node. \
"Always Ask" prompts on each command; "Always Allow" runs without prompts; \
"Never" disables `system.run`.
""")
.font(.footnote)
.foregroundStyle(.tertiary)
.fixedSize(horizontal: false, vertical: true)
}
SystemRunSettingsView()
VStack(alignment: .leading, spacing: 6) {
Text("Location Access")

View File

@ -1,81 +0,0 @@
import Foundation
import OSLog
enum MacNodeConfigFile {
private static let logger = Logger(subsystem: "com.clawdbot", category: "mac-node-config")
static func url() -> URL {
ClawdbotPaths.stateDirURL.appendingPathComponent("macos-node.json")
}
static func loadDict() -> [String: Any] {
let url = self.url()
guard FileManager.default.fileExists(atPath: url.path) else { return [:] }
do {
let data = try Data(contentsOf: url)
guard let root = try JSONSerialization.jsonObject(with: data) as? [String: Any] else {
self.logger.warning("mac node config JSON root invalid")
return [:]
}
return root
} catch {
self.logger.warning("mac node config read failed: \(error.localizedDescription, privacy: .public)")
return [:]
}
}
static func saveDict(_ dict: [String: Any]) {
do {
let data = try JSONSerialization.data(withJSONObject: dict, options: [.prettyPrinted, .sortedKeys])
let url = self.url()
try FileManager.default.createDirectory(
at: url.deletingLastPathComponent(),
withIntermediateDirectories: true)
try data.write(to: url, options: [.atomic])
try? FileManager.default.setAttributes([.posixPermissions: 0o600], ofItemAtPath: url.path)
} catch {
self.logger.error("mac node config save failed: \(error.localizedDescription, privacy: .public)")
}
}
static func systemRunPolicy() -> SystemRunPolicy? {
let root = self.loadDict()
let systemRun = root["systemRun"] as? [String: Any]
let raw = systemRun?["policy"] as? String
guard let raw, let policy = SystemRunPolicy(rawValue: raw) else { return nil }
return policy
}
static func setSystemRunPolicy(_ policy: SystemRunPolicy) {
var root = self.loadDict()
var systemRun = root["systemRun"] as? [String: Any] ?? [:]
systemRun["policy"] = policy.rawValue
root["systemRun"] = systemRun
self.saveDict(root)
}
static func systemRunAllowlist() -> [String]? {
let root = self.loadDict()
let systemRun = root["systemRun"] as? [String: Any]
return systemRun?["allowlist"] as? [String]
}
static func setSystemRunAllowlist(_ allowlist: [String]) {
let cleaned = allowlist
.map { $0.trimmingCharacters(in: .whitespacesAndNewlines) }
.filter { !$0.isEmpty }
var root = self.loadDict()
var systemRun = root["systemRun"] as? [String: Any] ?? [:]
if cleaned.isEmpty {
systemRun.removeValue(forKey: "allowlist")
} else {
systemRun["allowlist"] = cleaned
}
if systemRun.isEmpty {
root.removeValue(forKey: "systemRun")
} else {
root["systemRun"] = systemRun
}
self.saveDict(root)
}
}

View File

@ -256,6 +256,7 @@ final class AppDelegate: NSObject, NSApplicationDelegate {
}
TerminationSignalWatcher.shared.start()
NodePairingApprovalPrompter.shared.start()
ExecApprovalsPromptServer.shared.start()
MacNodeModeCoordinator.shared.start()
VoiceWakeGlobalSettingsSync.shared.start()
Task { PresenceReporter.shared.start() }
@ -280,6 +281,7 @@ final class AppDelegate: NSObject, NSApplicationDelegate {
func applicationWillTerminate(_ notification: Notification) {
PresenceReporter.shared.stop()
NodePairingApprovalPrompter.shared.stop()
ExecApprovalsPromptServer.shared.stop()
MacNodeModeCoordinator.shared.stop()
TerminationSignalWatcher.shared.stop()
VoiceWakeGlobalSettingsSync.shared.stop()

View File

@ -31,10 +31,10 @@ struct MenuContent: View {
self._updateStatus = Bindable(wrappedValue: updater?.updateStatus ?? UpdateStatus.disabled)
}
private var systemRunPolicyBinding: Binding<SystemRunPolicy> {
private var execApprovalModeBinding: Binding<ExecApprovalQuickMode> {
Binding(
get: { self.state.systemRunPolicy },
set: { self.state.systemRunPolicy = $0 })
get: { self.state.execApprovalMode },
set: { self.state.execApprovalMode = $0 })
}
var body: some View {
@ -74,12 +74,12 @@ struct MenuContent: View {
Toggle(isOn: self.$cameraEnabled) {
Label("Allow Camera", systemImage: "camera")
}
Picker(selection: self.systemRunPolicyBinding) {
ForEach(SystemRunPolicy.allCases) { policy in
Text(policy.title).tag(policy)
Picker(selection: self.execApprovalModeBinding) {
ForEach(ExecApprovalQuickMode.allCases) { mode in
Text(mode.title).tag(mode)
}
} label: {
Label("Node Run Commands", systemImage: "terminal")
Label("Exec Approvals", systemImage: "terminal")
}
Toggle(isOn: Binding(get: { self.state.canvasEnabled }, set: { self.state.canvasEnabled = $0 })) {
Label("Allow Canvas", systemImage: "rectangle.and.pencil.and.ellipsis")

View File

@ -14,6 +14,7 @@ final class MenuSessionsInjector: NSObject, NSMenuDelegate {
private weak var statusItem: NSStatusItem?
private var loadTask: Task<Void, Never>?
private var nodesLoadTask: Task<Void, Never>?
private var previewTasks: [Task<Void, Never>] = []
private var isMenuOpen = false
private var lastKnownMenuWidth: CGFloat?
private var menuOpenWidth: CGFloat?
@ -87,6 +88,7 @@ final class MenuSessionsInjector: NSObject, NSMenuDelegate {
self.menuOpenWidth = nil
self.loadTask?.cancel()
self.nodesLoadTask?.cancel()
self.cancelPreviewTasks()
}
func menuNeedsUpdate(_ menu: NSMenu) {
@ -107,6 +109,7 @@ extension MenuSessionsInjector {
private var mainSessionKey: String { WorkActivityStore.shared.mainSessionKey }
private func inject(into menu: NSMenu) {
self.cancelPreviewTasks()
// Remove any previous injected items.
for item in menu.items where item.tag == self.tag {
menu.removeItem(item)
@ -280,9 +283,7 @@ extension MenuSessionsInjector {
private func insertUsageSection(into menu: NSMenu, at cursor: Int, width: CGFloat) -> Int {
let rows = self.usageRows
let errorText = self.cachedUsageErrorText
if rows.isEmpty, errorText == nil {
if rows.isEmpty {
return cursor
}
@ -306,25 +307,6 @@ extension MenuSessionsInjector {
menu.insertItem(headerItem, at: cursor)
cursor += 1
if let errorText = errorText?.nonEmpty, !rows.isEmpty {
menu.insertItem(
self.makeMessageItem(
text: errorText,
symbolName: "exclamationmark.triangle",
width: width,
maxLines: 2),
at: cursor)
cursor += 1
}
if rows.isEmpty {
menu.insertItem(
self.makeMessageItem(text: errorText ?? "No usage available", symbolName: "minus", width: width),
at: cursor)
cursor += 1
return cursor
}
if let selectedProvider = self.selectedUsageProviderId,
let primary = rows.first(where: { $0.providerId.lowercased() == selectedProvider }),
rows.count > 1
@ -440,6 +422,8 @@ extension MenuSessionsInjector {
displayName: "Gateway",
platform: platform,
version: nil,
coreVersion: nil,
uiVersion: nil,
deviceFamily: nil,
modelIdentifier: nil,
remoteIp: host,
@ -473,15 +457,46 @@ extension MenuSessionsInjector {
item.tag = self.tag
item.isEnabled = false
let view = AnyView(SessionMenuPreviewView(
sessionKey: sessionKey,
width: width,
maxItems: 10,
maxLines: maxLines,
title: title))
item.view = self.makeHostedView(rootView: view, width: width, highlighted: false)
title: title,
items: [],
status: .loading))
let hosting = NSHostingView(rootView: view)
hosting.frame.size.width = max(1, width)
let size = hosting.fittingSize
hosting.frame = NSRect(origin: .zero, size: NSSize(width: width, height: size.height))
item.view = hosting
let task = Task { [weak hosting] in
let snapshot = await SessionMenuPreviewLoader.load(sessionKey: sessionKey, maxItems: 10)
guard !Task.isCancelled else { return }
await MainActor.run {
guard let hosting else { return }
let nextView = AnyView(SessionMenuPreviewView(
width: width,
maxLines: maxLines,
title: title,
items: snapshot.items,
status: snapshot.status))
hosting.rootView = nextView
hosting.invalidateIntrinsicContentSize()
hosting.frame.size.width = max(1, width)
let size = hosting.fittingSize
hosting.frame.size.height = size.height
}
}
self.previewTasks.append(task)
return item
}
private func cancelPreviewTasks() {
for task in self.previewTasks {
task.cancel()
}
self.previewTasks.removeAll()
}
private func makeMessageItem(text: String, symbolName: String, width: CGFloat, maxLines: Int? = 2) -> NSMenuItem {
let view = AnyView(
HStack(alignment: .top, spacing: 8) {
@ -559,14 +574,11 @@ extension MenuSessionsInjector {
do {
self.cachedUsageSummary = try await UsageLoader.loadSummary()
self.cachedUsageErrorText = nil
self.usageCacheUpdatedAt = Date()
} catch {
if self.cachedUsageSummary == nil {
self.cachedUsageErrorText = self.compactUsageError(error)
}
self.usageCacheUpdatedAt = Date()
self.cachedUsageSummary = nil
self.cachedUsageErrorText = nil
}
self.usageCacheUpdatedAt = Date()
}
private func compactUsageError(_ error: Error) -> String {
@ -747,8 +759,8 @@ extension MenuSessionsInjector {
menu.addItem(self.makeNodeCopyItem(label: "Platform", value: platform))
}
if let version = entry.version?.nonEmpty {
menu.addItem(self.makeNodeCopyItem(label: "Version", value: self.formatVersionLabel(version)))
if let version = NodeMenuEntryFormatter.detailRightVersion(entry)?.nonEmpty {
menu.addItem(self.makeNodeCopyItem(label: "Version", value: version))
}
menu.addItem(self.makeNodeDetailItem(label: "Connected", value: entry.isConnected ? "Yes" : "No"))

View File

@ -95,6 +95,8 @@ actor MacNodeBridgePairingClient {
displayName: hello.displayName,
platform: hello.platform,
version: hello.version,
coreVersion: hello.coreVersion,
uiVersion: hello.uiVersion,
deviceFamily: hello.deviceFamily,
modelIdentifier: hello.modelIdentifier,
caps: hello.caps,

View File

@ -43,7 +43,6 @@ final class MacNodeModeCoordinator {
private func run() async {
var retryDelay: UInt64 = 1_000_000_000
var lastCameraEnabled: Bool?
var lastSystemRunPolicy: SystemRunPolicy?
let defaults = UserDefaults.standard
while !Task.isCancelled {
if await MainActor.run(body: { AppStateStore.shared.isPaused }) {
@ -60,15 +59,6 @@ final class MacNodeModeCoordinator {
try? await Task.sleep(nanoseconds: 200_000_000)
}
let systemRunPolicy = SystemRunPolicy.load()
if lastSystemRunPolicy == nil {
lastSystemRunPolicy = systemRunPolicy
} else if lastSystemRunPolicy != systemRunPolicy {
lastSystemRunPolicy = systemRunPolicy
await self.session.disconnect()
try? await Task.sleep(nanoseconds: 200_000_000)
}
guard let target = await self.resolveBridgeEndpoint(timeoutSeconds: 5) else {
try? await Task.sleep(nanoseconds: min(retryDelay, 5_000_000_000))
retryDelay = min(retryDelay * 2, 10_000_000_000)
@ -89,8 +79,13 @@ final class MacNodeModeCoordinator {
if let mainSessionKey {
await self?.runtime.updateMainSessionKey(mainSessionKey)
}
await self?.runtime.setEventSender { [weak self] event, payload in
guard let self else { return }
try? await self.session.sendEvent(event: event, payloadJSON: payload)
}
},
onDisconnected: { reason in
onDisconnected: { [weak self] reason in
await self?.runtime.setEventSender(nil)
await MacNodeModeCoordinator.handleBridgeDisconnect(reason: reason)
},
onInvoke: { [weak self] req in
@ -119,12 +114,19 @@ final class MacNodeModeCoordinator {
let caps = self.currentCaps()
let commands = self.currentCommands(caps: caps)
let permissions = await self.currentPermissions()
let uiVersion = Bundle.main.object(forInfoDictionaryKey: "CFBundleShortVersionString") as? String
let liveGatewayVersion = await GatewayConnection.shared.cachedGatewayVersion()
let fallbackGatewayVersion = GatewayProcessManager.shared.environmentStatus.gatewayVersion
let coreVersion = (liveGatewayVersion ?? fallbackGatewayVersion)?
.trimmingCharacters(in: .whitespacesAndNewlines)
return BridgeHello(
nodeId: Self.nodeId(),
displayName: InstanceIdentity.displayName,
token: token,
platform: "macos",
version: Bundle.main.object(forInfoDictionaryKey: "CFBundleShortVersionString") as? String,
version: uiVersion,
coreVersion: coreVersion?.isEmpty == false ? coreVersion : nil,
uiVersion: uiVersion,
deviceFamily: "Mac",
modelIdentifier: InstanceIdentity.modelIdentifier,
caps: caps,
@ -161,13 +163,12 @@ final class MacNodeModeCoordinator {
ClawdbotCanvasA2UICommand.reset.rawValue,
MacNodeScreenCommand.record.rawValue,
ClawdbotSystemCommand.notify.rawValue,
ClawdbotSystemCommand.which.rawValue,
ClawdbotSystemCommand.run.rawValue,
ClawdbotSystemCommand.execApprovalsGet.rawValue,
ClawdbotSystemCommand.execApprovalsSet.rawValue,
]
if SystemRunPolicy.load() != .never {
commands.append(ClawdbotSystemCommand.which.rawValue)
commands.append(ClawdbotSystemCommand.run.rawValue)
}
let capsSet = Set(caps)
if capsSet.contains(ClawdbotCapability.camera.rawValue) {
commands.append(ClawdbotCameraCommand.list.rawValue)

View File

@ -8,6 +8,7 @@ actor MacNodeRuntime {
private let makeMainActorServices: () async -> any MacNodeRuntimeMainActorServices
private var cachedMainActorServices: (any MacNodeRuntimeMainActorServices)?
private var mainSessionKey: String = "main"
private var eventSender: (@Sendable (String, String?) async -> Void)?
init(
makeMainActorServices: @escaping () async -> any MacNodeRuntimeMainActorServices = {
@ -23,6 +24,10 @@ actor MacNodeRuntime {
self.mainSessionKey = trimmed
}
func setEventSender(_ sender: (@Sendable (String, String?) async -> Void)?) {
self.eventSender = sender
}
func handleInvoke(_ req: BridgeInvokeRequest) async -> BridgeInvokeResponse {
let command = req.command
if self.isCanvasCommand(command), !Self.canvasEnabled() {
@ -59,6 +64,10 @@ actor MacNodeRuntime {
return try await self.handleSystemWhich(req)
case ClawdbotSystemCommand.notify.rawValue:
return try await self.handleSystemNotify(req)
case ClawdbotSystemCommand.execApprovalsGet.rawValue:
return try await self.handleSystemExecApprovalsGet(req)
case ClawdbotSystemCommand.execApprovalsSet.rawValue:
return try await self.handleSystemExecApprovalsSet(req)
default:
return Self.errorResponse(req, code: .invalidRequest, message: "INVALID_REQUEST: unknown command")
}
@ -427,42 +436,168 @@ actor MacNodeRuntime {
guard !command.isEmpty else {
return Self.errorResponse(req, code: .invalidRequest, message: "INVALID_REQUEST: command required")
}
let displayCommand = ExecCommandFormatter.displayString(for: command, rawCommand: params.rawCommand)
let wasAllowlisted = SystemRunAllowlist.contains(command)
switch Self.systemRunPolicy() {
case .never:
let trimmedAgent = params.agentId?.trimmingCharacters(in: .whitespacesAndNewlines) ?? ""
let agentId = trimmedAgent.isEmpty ? nil : trimmedAgent
let approvals = ExecApprovalsStore.resolve(agentId: agentId)
let security = approvals.agent.security
let ask = approvals.agent.ask
let askFallback = approvals.agent.askFallback
let autoAllowSkills = approvals.agent.autoAllowSkills
let sessionKey = (params.sessionKey?.trimmingCharacters(in: .whitespacesAndNewlines).isEmpty == false)
? params.sessionKey!.trimmingCharacters(in: .whitespacesAndNewlines)
: self.mainSessionKey
let runId = UUID().uuidString
let env = Self.sanitizedEnv(params.env)
let resolution = ExecCommandResolution.resolve(
command: command,
rawCommand: params.rawCommand,
cwd: params.cwd,
env: env)
let allowlistMatch = security == .allowlist
? ExecAllowlistMatcher.match(entries: approvals.allowlist, resolution: resolution)
: nil
let skillAllow: Bool
if autoAllowSkills, let name = resolution?.executableName {
let bins = await SkillBinsCache.shared.currentBins()
skillAllow = bins.contains(name)
} else {
skillAllow = false
}
if security == .deny {
await self.emitExecEvent(
"exec.denied",
payload: ExecEventPayload(
sessionKey: sessionKey,
runId: runId,
host: "node",
command: displayCommand,
reason: "security=deny"))
return Self.errorResponse(
req,
code: .unavailable,
message: "SYSTEM_RUN_DISABLED: policy=never")
case .always:
break
case .ask:
if !wasAllowlisted {
let services = await self.mainActorServices()
let decision = await services.confirmSystemRun(
command: SystemRunAllowlist.displayString(for: command),
cwd: params.cwd)
switch decision {
case .allowOnce:
break
case .allowAlways:
SystemRunAllowlist.add(command)
case .deny:
message: "SYSTEM_RUN_DISABLED: security=deny")
}
let requiresAsk: Bool = {
if ask == .always { return true }
if ask == .onMiss && security == .allowlist && allowlistMatch == nil && !skillAllow { return true }
return false
}()
var approvedByAsk = false
if requiresAsk {
let decision: ExecApprovalDecision? = await ExecApprovalsPromptPresenter.prompt(
ExecApprovalPromptRequest(
command: displayCommand,
cwd: params.cwd,
host: "node",
security: security.rawValue,
ask: ask.rawValue,
agentId: agentId,
resolvedPath: resolution?.resolvedPath))
switch decision {
case .deny?:
await self.emitExecEvent(
"exec.denied",
payload: ExecEventPayload(
sessionKey: sessionKey,
runId: runId,
host: "node",
command: displayCommand,
reason: "user-denied"))
return Self.errorResponse(
req,
code: .unavailable,
message: "SYSTEM_RUN_DENIED: user denied")
case nil:
if askFallback == .full {
approvedByAsk = true
} else if askFallback == .allowlist {
if allowlistMatch != nil || skillAllow {
approvedByAsk = true
} else {
await self.emitExecEvent(
"exec.denied",
payload: ExecEventPayload(
sessionKey: sessionKey,
runId: runId,
host: "node",
command: displayCommand,
reason: "approval-required"))
return Self.errorResponse(
req,
code: .unavailable,
message: "SYSTEM_RUN_DENIED: approval required")
}
} else {
await self.emitExecEvent(
"exec.denied",
payload: ExecEventPayload(
sessionKey: sessionKey,
runId: runId,
host: "node",
command: displayCommand,
reason: "approval-required"))
return Self.errorResponse(
req,
code: .unavailable,
message: "SYSTEM_RUN_DENIED: user denied")
message: "SYSTEM_RUN_DENIED: approval required")
}
case .allowAlways?:
approvedByAsk = true
if security == .allowlist {
let pattern = resolution?.resolvedPath ??
resolution?.rawExecutable ??
command.first?.trimmingCharacters(in: .whitespacesAndNewlines) ??
""
if !pattern.isEmpty {
ExecApprovalsStore.addAllowlistEntry(agentId: agentId, pattern: pattern)
}
}
case .allowOnce?:
approvedByAsk = true
}
}
let env = Self.sanitizedEnv(params.env)
if security == .allowlist && allowlistMatch == nil && !skillAllow && !approvedByAsk {
await self.emitExecEvent(
"exec.denied",
payload: ExecEventPayload(
sessionKey: sessionKey,
runId: runId,
host: "node",
command: displayCommand,
reason: "allowlist-miss"))
return Self.errorResponse(
req,
code: .unavailable,
message: "SYSTEM_RUN_DENIED: allowlist miss")
}
if let match = allowlistMatch {
ExecApprovalsStore.recordAllowlistUse(
agentId: agentId,
pattern: match.pattern,
command: displayCommand,
resolvedPath: resolution?.resolvedPath)
}
if params.needsScreenRecording == true {
let authorized = await PermissionManager
.status([.screenRecording])[.screenRecording] ?? false
if !authorized {
await self.emitExecEvent(
"exec.denied",
payload: ExecEventPayload(
sessionKey: sessionKey,
runId: runId,
host: "node",
command: displayCommand,
reason: "permission:screenRecording"))
return Self.errorResponse(
req,
code: .unavailable,
@ -471,11 +606,33 @@ actor MacNodeRuntime {
}
let timeoutSec = params.timeoutMs.flatMap { Double($0) / 1000.0 }
await self.emitExecEvent(
"exec.started",
payload: ExecEventPayload(
sessionKey: sessionKey,
runId: runId,
host: "node",
command: displayCommand))
let result = await ShellExecutor.runDetailed(
command: command,
cwd: params.cwd,
env: env,
timeout: timeoutSec)
let combined = [result.stdout, result.stderr, result.errorMessage]
.compactMap { $0 }
.filter { !$0.isEmpty }
.joined(separator: "\n")
await self.emitExecEvent(
"exec.finished",
payload: ExecEventPayload(
sessionKey: sessionKey,
runId: runId,
host: "node",
command: displayCommand,
exitCode: result.exitCode,
timedOut: result.timedOut,
success: result.success,
output: ExecEventPayload.truncateOutput(combined)))
struct RunPayload: Encodable {
var exitCode: Int?
@ -523,6 +680,82 @@ actor MacNodeRuntime {
return BridgeInvokeResponse(id: req.id, ok: true, payloadJSON: payload)
}
private func handleSystemExecApprovalsGet(_ req: BridgeInvokeRequest) async throws -> BridgeInvokeResponse {
_ = ExecApprovalsStore.ensureFile()
let snapshot = ExecApprovalsStore.readSnapshot()
let redacted = ExecApprovalsSnapshot(
path: snapshot.path,
exists: snapshot.exists,
hash: snapshot.hash,
file: ExecApprovalsStore.redactForSnapshot(snapshot.file))
let payload = try Self.encodePayload(redacted)
return BridgeInvokeResponse(id: req.id, ok: true, payloadJSON: payload)
}
private func handleSystemExecApprovalsSet(_ req: BridgeInvokeRequest) async throws -> BridgeInvokeResponse {
struct SetParams: Decodable {
var file: ExecApprovalsFile
var baseHash: String?
}
let params = try Self.decodeParams(SetParams.self, from: req.paramsJSON)
let current = ExecApprovalsStore.ensureFile()
let snapshot = ExecApprovalsStore.readSnapshot()
if snapshot.exists {
if snapshot.hash.isEmpty {
return Self.errorResponse(
req,
code: .invalidRequest,
message: "INVALID_REQUEST: exec approvals base hash unavailable; reload and retry")
}
let baseHash = params.baseHash?.trimmingCharacters(in: .whitespacesAndNewlines) ?? ""
if baseHash.isEmpty {
return Self.errorResponse(
req,
code: .invalidRequest,
message: "INVALID_REQUEST: exec approvals base hash required; reload and retry")
}
if baseHash != snapshot.hash {
return Self.errorResponse(
req,
code: .invalidRequest,
message: "INVALID_REQUEST: exec approvals changed; reload and retry")
}
}
var normalized = ExecApprovalsStore.normalizeIncoming(params.file)
let socketPath = normalized.socket?.path?.trimmingCharacters(in: .whitespacesAndNewlines)
let token = normalized.socket?.token?.trimmingCharacters(in: .whitespacesAndNewlines)
let resolvedPath = (socketPath?.isEmpty == false)
? socketPath!
: current.socket?.path?.trimmingCharacters(in: .whitespacesAndNewlines) ??
ExecApprovalsStore.socketPath()
let resolvedToken = (token?.isEmpty == false)
? token!
: current.socket?.token?.trimmingCharacters(in: .whitespacesAndNewlines) ?? ""
normalized.socket = ExecApprovalsSocketConfig(path: resolvedPath, token: resolvedToken)
ExecApprovalsStore.saveFile(normalized)
let nextSnapshot = ExecApprovalsStore.readSnapshot()
let redacted = ExecApprovalsSnapshot(
path: nextSnapshot.path,
exists: nextSnapshot.exists,
hash: nextSnapshot.hash,
file: ExecApprovalsStore.redactForSnapshot(nextSnapshot.file))
let payload = try Self.encodePayload(redacted)
return BridgeInvokeResponse(id: req.id, ok: true, payloadJSON: payload)
}
private func emitExecEvent(_ event: String, payload: ExecEventPayload) async {
guard let sender = self.eventSender else { return }
guard let data = try? JSONEncoder().encode(payload),
let json = String(data: data, encoding: .utf8)
else {
return
}
await sender(event, json)
}
private func handleSystemNotify(_ req: BridgeInvokeRequest) async throws -> BridgeInvokeResponse {
let params = try Self.decodeParams(ClawdbotSystemNotifyParams.self, from: req.paramsJSON)
let title = params.title.trimmingCharacters(in: .whitespacesAndNewlines)
@ -589,10 +822,6 @@ actor MacNodeRuntime {
UserDefaults.standard.object(forKey: cameraEnabledKey) as? Bool ?? false
}
private nonisolated static func systemRunPolicy() -> SystemRunPolicy {
SystemRunPolicy.load()
}
private static let blockedEnvKeys: Set<String> = [
"PATH",
"NODE_OPTIONS",

View File

@ -1,14 +1,7 @@
import AppKit
import ClawdbotKit
import CoreLocation
import Foundation
enum SystemRunDecision: Sendable {
case allowOnce
case allowAlways
case deny
}
@MainActor
protocol MacNodeRuntimeMainActorServices: Sendable {
func recordScreen(
@ -24,8 +17,6 @@ protocol MacNodeRuntimeMainActorServices: Sendable {
desiredAccuracy: ClawdbotLocationAccuracy,
maxAgeMs: Int?,
timeoutMs: Int?) async throws -> CLLocation
func confirmSystemRun(command: String, cwd: String?) async -> SystemRunDecision
}
@MainActor
@ -67,30 +58,4 @@ final class LiveMacNodeRuntimeMainActorServices: MacNodeRuntimeMainActorServices
timeoutMs: timeoutMs)
}
func confirmSystemRun(command: String, cwd: String?) async -> SystemRunDecision {
let alert = NSAlert()
alert.alertStyle = .warning
alert.messageText = "Allow this command?"
var details = "Clawdbot wants to run:\n\n\(command)"
let trimmedCwd = cwd?.trimmingCharacters(in: .whitespacesAndNewlines) ?? ""
if !trimmedCwd.isEmpty {
details += "\n\nWorking directory:\n\(trimmedCwd)"
}
details += "\n\nThis runs on this Mac via node mode."
alert.informativeText = details
alert.addButton(withTitle: "Allow Once")
alert.addButton(withTitle: "Always Allow")
alert.addButton(withTitle: "Don't Allow")
switch alert.runModal() {
case .alertFirstButtonReturn:
return .allowOnce
case .alertSecondButtonReturn:
return .allowAlways
default:
return .deny
}
}
}

View File

@ -35,8 +35,9 @@ struct NodeMenuEntryFormatter {
if let platform = self.platformText(entry) {
parts.append("platform \(platform)")
}
if let version = entry.version?.nonEmpty {
parts.append("app \(self.compactVersion(version))")
let versionLabels = self.versionLabels(entry)
if !versionLabels.isEmpty {
parts.append(versionLabels.joined(separator: " · "))
}
parts.append("status \(self.roleText(entry))")
return parts.joined(separator: " · ")
@ -60,8 +61,9 @@ struct NodeMenuEntryFormatter {
}
static func detailRightVersion(_ entry: NodeInfo) -> String? {
guard let version = entry.version?.nonEmpty else { return nil }
return self.shortVersionLabel(version)
let labels = self.versionLabels(entry, compact: false)
if labels.isEmpty { return nil }
return labels.joined(separator: " · ")
}
static func platformText(_ entry: NodeInfo) -> String? {
@ -127,6 +129,39 @@ struct NodeMenuEntryFormatter {
return compact
}
private static func versionLabels(_ entry: NodeInfo, compact: Bool = true) -> [String] {
let (core, ui) = self.resolveVersions(entry)
var labels: [String] = []
if let core {
let label = compact ? self.compactVersion(core) : self.shortVersionLabel(core)
labels.append("core \(label)")
}
if let ui {
let label = compact ? self.compactVersion(ui) : self.shortVersionLabel(ui)
labels.append("ui \(label)")
}
return labels
}
private static func resolveVersions(_ entry: NodeInfo) -> (core: String?, ui: String?) {
let core = entry.coreVersion?.nonEmpty
let ui = entry.uiVersion?.nonEmpty
if core != nil || ui != nil {
return (core, ui)
}
guard let legacy = entry.version?.nonEmpty else { return (nil, nil) }
if self.isHeadlessPlatform(entry) {
return (legacy, nil)
}
return (nil, legacy)
}
private static func isHeadlessPlatform(_ entry: NodeInfo) -> Bool {
let raw = entry.platform?.trimmingCharacters(in: .whitespacesAndNewlines).lowercased() ?? ""
if raw == "darwin" || raw == "linux" || raw == "win32" || raw == "windows" { return true }
return false
}
static func leadingSymbol(_ entry: NodeInfo) -> String {
if self.isGateway(entry) {
return self.safeSystemSymbol(

View File

@ -7,6 +7,8 @@ struct NodeInfo: Identifiable, Codable {
let displayName: String?
let platform: String?
let version: String?
let coreVersion: String?
let uiVersion: String?
let deviceFamily: String?
let modelIdentifier: String?
let remoteIp: String?

View File

@ -3,13 +3,13 @@ import ClawdbotKit
import OSLog
import SwiftUI
private struct SessionPreviewItem: Identifiable, Sendable {
struct SessionPreviewItem: Identifiable, Sendable {
let id: String
let role: PreviewRole
let text: String
}
private enum PreviewRole: String, Sendable {
enum PreviewRole: String, Sendable {
case user
case assistant
case tool
@ -27,7 +27,7 @@ private enum PreviewRole: String, Sendable {
}
}
private actor SessionPreviewCache {
actor SessionPreviewCache {
static let shared = SessionPreviewCache()
private struct CacheEntry {
@ -52,25 +52,33 @@ private actor SessionPreviewCache {
}
}
struct SessionMenuPreviewView: View {
private static let logger = Logger(subsystem: "com.clawdbot", category: "SessionPreview")
private static let previewTimeoutSeconds: Double = 4
let sessionKey: String
let width: CGFloat
let maxItems: Int
let maxLines: Int
let title: String
@Environment(\.menuItemHighlighted) private var isHighlighted
@State private var items: [SessionPreviewItem] = []
@State private var status: LoadStatus = .loading
private struct PreviewTimeoutError: LocalizedError {
var errorDescription: String? { "preview timeout" }
#if DEBUG
extension SessionPreviewCache {
func _testSet(items: [SessionPreviewItem], for sessionKey: String, updatedAt: Date = Date()) {
self.entries[sessionKey] = CacheEntry(items: items, updatedAt: updatedAt)
}
private enum LoadStatus: Equatable {
func _testReset() {
self.entries = [:]
}
}
#endif
struct SessionMenuPreviewSnapshot: Sendable {
let items: [SessionPreviewItem]
let status: SessionMenuPreviewView.LoadStatus
}
struct SessionMenuPreviewView: View {
let width: CGFloat
let maxLines: Int
let title: String
let items: [SessionPreviewItem]
let status: LoadStatus
@Environment(\.menuItemHighlighted) private var isHighlighted
enum LoadStatus: Equatable {
case loading
case ready
case empty
@ -85,10 +93,6 @@ struct SessionMenuPreviewView: View {
self.isHighlighted ? Color(nsColor: .selectedMenuItemTextColor).opacity(0.85) : .secondary
}
private var previewLimit: Int {
min(max(self.maxItems * 3, 20), 120)
}
var body: some View {
VStack(alignment: .leading, spacing: 8) {
HStack(alignment: .firstTextBaseline, spacing: 4) {
@ -123,9 +127,6 @@ struct SessionMenuPreviewView: View {
.padding(.leading, 16)
.padding(.trailing, 11)
.frame(width: max(1, self.width), alignment: .leading)
.task(id: self.sessionKey) {
await self.loadPreview()
}
}
@ViewBuilder
@ -157,55 +158,59 @@ struct SessionMenuPreviewView: View {
}
}
private func loadPreview() async {
if let cached = await SessionPreviewCache.shared.cachedItems(for: self.sessionKey, maxAge: 12) {
await MainActor.run {
self.items = cached
self.status = cached.isEmpty ? .empty : .ready
}
return
}
}
await MainActor.run {
self.status = .loading
enum SessionMenuPreviewLoader {
private static let logger = Logger(subsystem: "com.clawdbot", category: "SessionPreview")
private static let previewTimeoutSeconds: Double = 4
private static let cacheMaxAgeSeconds: TimeInterval = 30
private struct PreviewTimeoutError: LocalizedError {
var errorDescription: String? { "preview timeout" }
}
static func load(sessionKey: String, maxItems: Int) async -> SessionMenuPreviewSnapshot {
if let cached = await SessionPreviewCache.shared.cachedItems(for: sessionKey, maxAge: cacheMaxAgeSeconds) {
return Self.snapshot(from: cached)
}
do {
let timeoutMs = Int(Self.previewTimeoutSeconds * 1000)
let timeoutMs = Int(self.previewTimeoutSeconds * 1000)
let payload = try await AsyncTimeout.withTimeout(
seconds: Self.previewTimeoutSeconds,
seconds: self.previewTimeoutSeconds,
onTimeout: { PreviewTimeoutError() },
operation: {
try await GatewayConnection.shared.chatHistory(
sessionKey: self.sessionKey,
limit: self.previewLimit,
sessionKey: sessionKey,
limit: self.previewLimit(for: maxItems),
timeoutMs: timeoutMs)
})
let built = Self.previewItems(from: payload, maxItems: self.maxItems)
await SessionPreviewCache.shared.store(items: built, for: self.sessionKey)
await MainActor.run {
self.items = built
self.status = built.isEmpty ? .empty : .ready
}
let built = Self.previewItems(from: payload, maxItems: maxItems)
await SessionPreviewCache.shared.store(items: built, for: sessionKey)
return Self.snapshot(from: built)
} catch is CancellationError {
return
return SessionMenuPreviewSnapshot(items: [], status: .loading)
} catch {
let fallback = await SessionPreviewCache.shared.lastItems(for: self.sessionKey)
await MainActor.run {
if let fallback {
self.items = fallback
self.status = fallback.isEmpty ? .empty : .ready
} else {
self.status = .error("Preview unavailable")
}
let fallback = await SessionPreviewCache.shared.lastItems(for: sessionKey)
if let fallback {
return Self.snapshot(from: fallback)
}
let errorDescription = String(describing: error)
Self.logger.warning(
"Session preview failed session=\(self.sessionKey, privacy: .public) " +
"Session preview failed session=\(sessionKey, privacy: .public) " +
"error=\(errorDescription, privacy: .public)")
return SessionMenuPreviewSnapshot(items: [], status: .error("Preview unavailable"))
}
}
private static func snapshot(from items: [SessionPreviewItem]) -> SessionMenuPreviewSnapshot {
SessionMenuPreviewSnapshot(items: items, status: items.isEmpty ? .empty : .ready)
}
private static func previewLimit(for maxItems: Int) -> Int {
min(max(maxItems * 3, 20), 120)
}
private static func previewItems(
from payload: ClawdbotChatHistoryPayload,
maxItems: Int) -> [SessionPreviewItem]

View File

@ -1,89 +0,0 @@
import Foundation
enum SystemRunPolicy: String, CaseIterable, Identifiable {
case never
case ask
case always
var id: String { self.rawValue }
var title: String {
switch self {
case .never:
"Never"
case .ask:
"Always Ask"
case .always:
"Always Allow"
}
}
static func load(from defaults: UserDefaults = .standard) -> SystemRunPolicy {
if let policy = MacNodeConfigFile.systemRunPolicy() {
return policy
}
if let raw = defaults.string(forKey: systemRunPolicyKey),
let policy = SystemRunPolicy(rawValue: raw)
{
MacNodeConfigFile.setSystemRunPolicy(policy)
return policy
}
if let legacy = defaults.object(forKey: systemRunEnabledKey) as? Bool {
let policy: SystemRunPolicy = legacy ? .ask : .never
MacNodeConfigFile.setSystemRunPolicy(policy)
return policy
}
let fallback: SystemRunPolicy = .ask
MacNodeConfigFile.setSystemRunPolicy(fallback)
return fallback
}
}
enum SystemRunAllowlist {
static func key(for argv: [String]) -> String {
let trimmed = argv.map { $0.trimmingCharacters(in: .whitespacesAndNewlines) }
guard !trimmed.isEmpty else { return "" }
if let data = try? JSONEncoder().encode(trimmed),
let json = String(data: data, encoding: .utf8)
{
return json
}
return trimmed.joined(separator: " ")
}
static func displayString(for argv: [String]) -> String {
argv.map { arg in
let trimmed = arg.trimmingCharacters(in: .whitespacesAndNewlines)
guard !trimmed.isEmpty else { return "\"\"" }
let needsQuotes = trimmed.contains { $0.isWhitespace || $0 == "\"" }
if !needsQuotes { return trimmed }
let escaped = trimmed.replacingOccurrences(of: "\"", with: "\\\"")
return "\"\(escaped)\""
}.joined(separator: " ")
}
static func load(from defaults: UserDefaults = .standard) -> Set<String> {
if let allowlist = MacNodeConfigFile.systemRunAllowlist() {
return Set(allowlist)
}
if let legacy = defaults.stringArray(forKey: systemRunAllowlistKey), !legacy.isEmpty {
MacNodeConfigFile.setSystemRunAllowlist(legacy)
return Set(legacy)
}
return []
}
static func contains(_ argv: [String], defaults: UserDefaults = .standard) -> Bool {
let key = key(for: argv)
return self.load(from: defaults).contains(key)
}
static func add(_ argv: [String], defaults: UserDefaults = .standard) {
let key = key(for: argv)
guard !key.isEmpty else { return }
var allowlist = self.load(from: defaults)
if allowlist.insert(key).inserted {
MacNodeConfigFile.setSystemRunAllowlist(Array(allowlist).sorted())
}
}
}

View File

@ -0,0 +1,401 @@
import Foundation
import Observation
import SwiftUI
struct SystemRunSettingsView: View {
@State private var model = ExecApprovalsSettingsModel()
@State private var tab: ExecApprovalsSettingsTab = .policy
@State private var newPattern: String = ""
var body: some View {
VStack(alignment: .leading, spacing: 8) {
HStack(alignment: .center, spacing: 12) {
Text("Exec approvals")
.font(.body)
Spacer(minLength: 0)
Picker("Agent", selection: Binding(
get: { self.model.selectedAgentId },
set: { self.model.selectAgent($0) }))
{
ForEach(self.model.agentPickerIds, id: \.self) { id in
Text(self.model.label(for: id)).tag(id)
}
}
.pickerStyle(.menu)
.frame(width: 180, alignment: .trailing)
}
Picker("", selection: self.$tab) {
ForEach(ExecApprovalsSettingsTab.allCases) { tab in
Text(tab.title).tag(tab)
}
}
.pickerStyle(.segmented)
.frame(width: 320)
if self.tab == .policy {
self.policyView
} else {
self.allowlistView
}
}
.task { await self.model.refresh() }
.onChange(of: self.tab) { _, _ in
Task { await self.model.refreshSkillBins() }
}
}
private var policyView: some View {
VStack(alignment: .leading, spacing: 8) {
Picker("", selection: Binding(
get: { self.model.security },
set: { self.model.setSecurity($0) }))
{
ForEach(ExecSecurity.allCases) { security in
Text(security.title).tag(security)
}
}
.labelsHidden()
.pickerStyle(.menu)
Picker("", selection: Binding(
get: { self.model.ask },
set: { self.model.setAsk($0) }))
{
ForEach(ExecAsk.allCases) { ask in
Text(ask.title).tag(ask)
}
}
.labelsHidden()
.pickerStyle(.menu)
Picker("", selection: Binding(
get: { self.model.askFallback },
set: { self.model.setAskFallback($0) }))
{
ForEach(ExecSecurity.allCases) { mode in
Text("Fallback: \(mode.title)").tag(mode)
}
}
.labelsHidden()
.pickerStyle(.menu)
Text(self.model.isDefaultsScope
? "Defaults apply when an agent has no overrides. Ask controls prompt behavior; fallback is used when no companion UI is reachable."
: "Security controls whether system.run can execute on this Mac when paired as a node. Ask controls prompt behavior; fallback is used when no companion UI is reachable.")
.font(.footnote)
.foregroundStyle(.tertiary)
.fixedSize(horizontal: false, vertical: true)
}
}
private var allowlistView: some View {
VStack(alignment: .leading, spacing: 10) {
Toggle("Auto-allow skill CLIs", isOn: Binding(
get: { self.model.autoAllowSkills },
set: { self.model.setAutoAllowSkills($0) }))
if self.model.autoAllowSkills, !self.model.skillBins.isEmpty {
Text("Skill CLIs: \(self.model.skillBins.joined(separator: ", "))")
.font(.footnote)
.foregroundStyle(.secondary)
}
if self.model.isDefaultsScope {
Text("Allowlists are per-agent. Select an agent to edit its allowlist.")
.font(.footnote)
.foregroundStyle(.secondary)
} else {
HStack(spacing: 8) {
TextField("Add allowlist pattern (case-insensitive globs)", text: self.$newPattern)
.textFieldStyle(.roundedBorder)
Button("Add") {
let pattern = self.newPattern.trimmingCharacters(in: .whitespacesAndNewlines)
guard !pattern.isEmpty else { return }
self.model.addEntry(pattern)
self.newPattern = ""
}
.buttonStyle(.bordered)
.disabled(self.newPattern.trimmingCharacters(in: .whitespacesAndNewlines).isEmpty)
}
if self.model.entries.isEmpty {
Text("No allowlisted commands yet.")
.font(.footnote)
.foregroundStyle(.secondary)
} else {
VStack(alignment: .leading, spacing: 8) {
ForEach(Array(self.model.entries.enumerated()), id: \.offset) { index, _ in
ExecAllowlistRow(
entry: Binding(
get: { self.model.entries[index] },
set: { self.model.updateEntry($0, at: index) }),
onRemove: { self.model.removeEntry(at: index) })
}
}
}
}
}
}
}
private enum ExecApprovalsSettingsTab: String, CaseIterable, Identifiable {
case policy
case allowlist
var id: String { self.rawValue }
var title: String {
switch self {
case .policy: "Access"
case .allowlist: "Allowlist"
}
}
}
struct ExecAllowlistRow: View {
@Binding var entry: ExecAllowlistEntry
let onRemove: () -> Void
@State private var draftPattern: String = ""
private static let relativeFormatter: RelativeDateTimeFormatter = {
let formatter = RelativeDateTimeFormatter()
formatter.unitsStyle = .short
return formatter
}()
var body: some View {
VStack(alignment: .leading, spacing: 4) {
HStack(spacing: 8) {
TextField("Pattern", text: self.patternBinding)
.textFieldStyle(.roundedBorder)
Button(role: .destructive) {
self.onRemove()
} label: {
Image(systemName: "trash")
}
.buttonStyle(.borderless)
}
if let lastUsedAt = self.entry.lastUsedAt {
let date = Date(timeIntervalSince1970: lastUsedAt / 1000.0)
Text("Last used \(Self.relativeFormatter.localizedString(for: date, relativeTo: Date()))")
.font(.caption)
.foregroundStyle(.secondary)
}
if let lastUsedCommand = self.entry.lastUsedCommand, !lastUsedCommand.isEmpty {
Text("Last command: \(lastUsedCommand)")
.font(.caption)
.foregroundStyle(.secondary)
}
if let lastResolvedPath = self.entry.lastResolvedPath, !lastResolvedPath.isEmpty {
Text("Resolved path: \(lastResolvedPath)")
.font(.caption)
.foregroundStyle(.secondary)
}
}
.onAppear {
self.draftPattern = self.entry.pattern
}
}
private var patternBinding: Binding<String> {
Binding(
get: { self.draftPattern.isEmpty ? self.entry.pattern : self.draftPattern },
set: { newValue in
self.draftPattern = newValue
self.entry.pattern = newValue
})
}
}
@MainActor
@Observable
final class ExecApprovalsSettingsModel {
private static let defaultsScopeId = "__defaults__"
var agentIds: [String] = []
var selectedAgentId: String = "main"
var defaultAgentId: String = "main"
var security: ExecSecurity = .deny
var ask: ExecAsk = .onMiss
var askFallback: ExecSecurity = .deny
var autoAllowSkills = false
var entries: [ExecAllowlistEntry] = []
var skillBins: [String] = []
var agentPickerIds: [String] {
[Self.defaultsScopeId] + self.agentIds
}
var isDefaultsScope: Bool {
self.selectedAgentId == Self.defaultsScopeId
}
func label(for id: String) -> String {
if id == Self.defaultsScopeId { return "Defaults" }
return id
}
func refresh() async {
await self.refreshAgents()
self.loadSettings(for: self.selectedAgentId)
await self.refreshSkillBins()
}
func refreshAgents() async {
let root = await ConfigStore.load()
let agents = root["agents"] as? [String: Any]
let list = agents?["list"] as? [[String: Any]] ?? []
var ids: [String] = []
var seen = Set<String>()
var defaultId: String?
for entry in list {
guard let raw = entry["id"] as? String else { continue }
let trimmed = raw.trimmingCharacters(in: .whitespacesAndNewlines)
guard !trimmed.isEmpty else { continue }
if !seen.insert(trimmed).inserted { continue }
ids.append(trimmed)
if (entry["default"] as? Bool) == true, defaultId == nil {
defaultId = trimmed
}
}
if ids.isEmpty {
ids = ["main"]
defaultId = "main"
} else if defaultId == nil {
defaultId = ids.first
}
self.agentIds = ids
self.defaultAgentId = defaultId ?? "main"
if self.selectedAgentId == Self.defaultsScopeId {
return
}
if !self.agentIds.contains(self.selectedAgentId) {
self.selectedAgentId = self.defaultAgentId
}
}
func selectAgent(_ id: String) {
self.selectedAgentId = id
self.loadSettings(for: id)
Task { await self.refreshSkillBins() }
}
func loadSettings(for agentId: String) {
if agentId == Self.defaultsScopeId {
let defaults = ExecApprovalsStore.resolveDefaults()
self.security = defaults.security
self.ask = defaults.ask
self.askFallback = defaults.askFallback
self.autoAllowSkills = defaults.autoAllowSkills
self.entries = []
return
}
let resolved = ExecApprovalsStore.resolve(agentId: agentId)
self.security = resolved.agent.security
self.ask = resolved.agent.ask
self.askFallback = resolved.agent.askFallback
self.autoAllowSkills = resolved.agent.autoAllowSkills
self.entries = resolved.allowlist
.sorted { $0.pattern.localizedCaseInsensitiveCompare($1.pattern) == .orderedAscending }
}
func setSecurity(_ security: ExecSecurity) {
self.security = security
if self.isDefaultsScope {
ExecApprovalsStore.updateDefaults { defaults in
defaults.security = security
}
} else {
ExecApprovalsStore.updateAgentSettings(agentId: self.selectedAgentId) { entry in
entry.security = security
}
}
self.syncQuickMode()
}
func setAsk(_ ask: ExecAsk) {
self.ask = ask
if self.isDefaultsScope {
ExecApprovalsStore.updateDefaults { defaults in
defaults.ask = ask
}
} else {
ExecApprovalsStore.updateAgentSettings(agentId: self.selectedAgentId) { entry in
entry.ask = ask
}
}
self.syncQuickMode()
}
func setAskFallback(_ mode: ExecSecurity) {
self.askFallback = mode
if self.isDefaultsScope {
ExecApprovalsStore.updateDefaults { defaults in
defaults.askFallback = mode
}
} else {
ExecApprovalsStore.updateAgentSettings(agentId: self.selectedAgentId) { entry in
entry.askFallback = mode
}
}
}
func setAutoAllowSkills(_ enabled: Bool) {
self.autoAllowSkills = enabled
if self.isDefaultsScope {
ExecApprovalsStore.updateDefaults { defaults in
defaults.autoAllowSkills = enabled
}
} else {
ExecApprovalsStore.updateAgentSettings(agentId: self.selectedAgentId) { entry in
entry.autoAllowSkills = enabled
}
}
Task { await self.refreshSkillBins(force: enabled) }
}
func addEntry(_ pattern: String) {
guard !self.isDefaultsScope else { return }
let trimmed = pattern.trimmingCharacters(in: .whitespacesAndNewlines)
guard !trimmed.isEmpty else { return }
self.entries.append(ExecAllowlistEntry(pattern: trimmed, lastUsedAt: nil))
ExecApprovalsStore.updateAllowlist(agentId: self.selectedAgentId, allowlist: self.entries)
}
func updateEntry(_ entry: ExecAllowlistEntry, at index: Int) {
guard !self.isDefaultsScope else { return }
guard self.entries.indices.contains(index) else { return }
self.entries[index] = entry
ExecApprovalsStore.updateAllowlist(agentId: self.selectedAgentId, allowlist: self.entries)
}
func removeEntry(at index: Int) {
guard !self.isDefaultsScope else { return }
guard self.entries.indices.contains(index) else { return }
self.entries.remove(at: index)
ExecApprovalsStore.updateAllowlist(agentId: self.selectedAgentId, allowlist: self.entries)
}
func refreshSkillBins(force: Bool = false) async {
guard self.autoAllowSkills else {
self.skillBins = []
return
}
let bins = await SkillBinsCache.shared.currentBins(force: force)
self.skillBins = bins.sorted()
}
private func syncQuickMode() {
if self.isDefaultsScope {
AppStateStore.shared.execApprovalMode = ExecApprovalQuickMode.from(security: self.security, ask: self.ask)
return
}
if self.selectedAgentId == self.defaultAgentId || self.agentIds.count <= 1 {
AppStateStore.shared.execApprovalMode = ExecApprovalQuickMode.from(security: self.security, ask: self.ask)
}
}
}

View File

@ -530,6 +530,8 @@ public struct NodePairRequestParams: Codable, Sendable {
public let displayname: String?
public let platform: String?
public let version: String?
public let coreversion: String?
public let uiversion: String?
public let devicefamily: String?
public let modelidentifier: String?
public let caps: [String]?
@ -542,6 +544,8 @@ public struct NodePairRequestParams: Codable, Sendable {
displayname: String?,
platform: String?,
version: String?,
coreversion: String?,
uiversion: String?,
devicefamily: String?,
modelidentifier: String?,
caps: [String]?,
@ -553,6 +557,8 @@ public struct NodePairRequestParams: Codable, Sendable {
self.displayname = displayname
self.platform = platform
self.version = version
self.coreversion = coreversion
self.uiversion = uiversion
self.devicefamily = devicefamily
self.modelidentifier = modelidentifier
self.caps = caps
@ -565,6 +571,8 @@ public struct NodePairRequestParams: Codable, Sendable {
case displayname = "displayName"
case platform
case version
case coreversion = "coreVersion"
case uiversion = "uiVersion"
case devicefamily = "deviceFamily"
case modelidentifier = "modelIdentifier"
case caps
@ -760,6 +768,10 @@ public struct SessionsPatchParams: Codable, Sendable {
public let reasoninglevel: AnyCodable?
public let responseusage: AnyCodable?
public let elevatedlevel: AnyCodable?
public let exechost: AnyCodable?
public let execsecurity: AnyCodable?
public let execask: AnyCodable?
public let execnode: AnyCodable?
public let model: AnyCodable?
public let spawnedby: AnyCodable?
public let sendpolicy: AnyCodable?
@ -773,6 +785,10 @@ public struct SessionsPatchParams: Codable, Sendable {
reasoninglevel: AnyCodable?,
responseusage: AnyCodable?,
elevatedlevel: AnyCodable?,
exechost: AnyCodable?,
execsecurity: AnyCodable?,
execask: AnyCodable?,
execnode: AnyCodable?,
model: AnyCodable?,
spawnedby: AnyCodable?,
sendpolicy: AnyCodable?,
@ -785,6 +801,10 @@ public struct SessionsPatchParams: Codable, Sendable {
self.reasoninglevel = reasoninglevel
self.responseusage = responseusage
self.elevatedlevel = elevatedlevel
self.exechost = exechost
self.execsecurity = execsecurity
self.execask = execask
self.execnode = execnode
self.model = model
self.spawnedby = spawnedby
self.sendpolicy = sendpolicy
@ -798,6 +818,10 @@ public struct SessionsPatchParams: Codable, Sendable {
case reasoninglevel = "reasoningLevel"
case responseusage = "responseUsage"
case elevatedlevel = "elevatedLevel"
case exechost = "execHost"
case execsecurity = "execSecurity"
case execask = "execAsk"
case execnode = "execNode"
case model
case spawnedby = "spawnedBy"
case sendpolicy = "sendPolicy"
@ -1616,6 +1640,85 @@ public struct LogsTailResult: Codable, Sendable {
}
}
public struct ExecApprovalsGetParams: Codable, Sendable {
}
public struct ExecApprovalsSetParams: Codable, Sendable {
public let file: [String: AnyCodable]
public let basehash: String?
public init(
file: [String: AnyCodable],
basehash: String?
) {
self.file = file
self.basehash = basehash
}
private enum CodingKeys: String, CodingKey {
case file
case basehash = "baseHash"
}
}
public struct ExecApprovalsNodeGetParams: Codable, Sendable {
public let nodeid: String
public init(
nodeid: String
) {
self.nodeid = nodeid
}
private enum CodingKeys: String, CodingKey {
case nodeid = "nodeId"
}
}
public struct ExecApprovalsNodeSetParams: Codable, Sendable {
public let nodeid: String
public let file: [String: AnyCodable]
public let basehash: String?
public init(
nodeid: String,
file: [String: AnyCodable],
basehash: String?
) {
self.nodeid = nodeid
self.file = file
self.basehash = basehash
}
private enum CodingKeys: String, CodingKey {
case nodeid = "nodeId"
case file
case basehash = "baseHash"
}
}
public struct ExecApprovalsSnapshot: Codable, Sendable {
public let path: String
public let exists: Bool
public let hash: String
public let file: [String: AnyCodable]
public init(
path: String,
exists: Bool,
hash: String,
file: [String: AnyCodable]
) {
self.path = path
self.exists = exists
self.hash = hash
self.file = file
}
private enum CodingKeys: String, CodingKey {
case path
case exists
case hash
case file
}
}
public struct ChatHistoryParams: Codable, Sendable {
public let sessionkey: String
public let limit: Int?

View File

@ -34,7 +34,7 @@ import Testing
let clawdbotPath = tmp.appendingPathComponent("node_modules/.bin/clawdbot")
try self.makeExec(at: clawdbotPath)
let cmd = CommandResolver.clawdbotCommand(subcommand: "gateway", defaults: defaults)
let cmd = CommandResolver.clawdbotCommand(subcommand: "gateway", defaults: defaults, configRoot: [:])
#expect(cmd.prefix(2).elementsEqual([clawdbotPath.path, "gateway"]))
}
@ -55,6 +55,7 @@ import Testing
let cmd = CommandResolver.clawdbotCommand(
subcommand: "rpc",
defaults: defaults,
configRoot: [:],
searchPaths: [tmp.appendingPathComponent("node_modules/.bin").path])
#expect(cmd.count >= 3)
@ -75,7 +76,7 @@ import Testing
let pnpmPath = tmp.appendingPathComponent("node_modules/.bin/pnpm")
try self.makeExec(at: pnpmPath)
let cmd = CommandResolver.clawdbotCommand(subcommand: "rpc", defaults: defaults)
let cmd = CommandResolver.clawdbotCommand(subcommand: "rpc", defaults: defaults, configRoot: [:])
#expect(cmd.prefix(4).elementsEqual([pnpmPath.path, "--silent", "clawdbot", "rpc"]))
}
@ -93,7 +94,8 @@ import Testing
let cmd = CommandResolver.clawdbotCommand(
subcommand: "health",
extraArgs: ["--json", "--timeout", "5"],
defaults: defaults)
defaults: defaults,
configRoot: [:])
#expect(cmd.prefix(5).elementsEqual([pnpmPath.path, "--silent", "clawdbot", "health", "--json"]))
#expect(cmd.suffix(2).elementsEqual(["--timeout", "5"]))
@ -114,7 +116,11 @@ import Testing
defaults.set("/tmp/id_ed25519", forKey: remoteIdentityKey)
defaults.set("/srv/clawdbot", forKey: remoteProjectRootKey)
let cmd = CommandResolver.clawdbotCommand(subcommand: "status", extraArgs: ["--json"], defaults: defaults)
let cmd = CommandResolver.clawdbotCommand(
subcommand: "status",
extraArgs: ["--json"],
defaults: defaults,
configRoot: [:])
#expect(cmd.first == "/usr/bin/ssh")
#expect(cmd.contains("clawd@example.com"))
@ -128,4 +134,27 @@ import Testing
#expect(script.contains("CLI="))
}
}
@Test func configRootLocalOverridesRemoteDefaults() async throws {
let defaults = self.makeDefaults()
defaults.set(AppState.ConnectionMode.remote.rawValue, forKey: connectionModeKey)
defaults.set("clawd@example.com:2222", forKey: remoteTargetKey)
let tmp = try makeTempDir()
CommandResolver.setProjectRoot(tmp.path)
let clawdbotPath = tmp.appendingPathComponent("node_modules/.bin/clawdbot")
try self.makeExec(at: clawdbotPath)
let cmd = CommandResolver.clawdbotCommand(
subcommand: "daemon",
defaults: defaults,
configRoot: ["gateway": ["mode": "local"]])
#expect(cmd.first == clawdbotPath.path)
#expect(cmd.count >= 2)
if cmd.count >= 2 {
#expect(cmd[1] == "daemon")
}
}
}

View File

@ -0,0 +1,49 @@
import Foundation
import Testing
@testable import Clawdbot
struct ExecAllowlistTests {
@Test func matchUsesResolvedPath() {
let entry = ExecAllowlistEntry(pattern: "/opt/homebrew/bin/rg")
let resolution = ExecCommandResolution(
rawExecutable: "rg",
resolvedPath: "/opt/homebrew/bin/rg",
executableName: "rg",
cwd: nil)
let match = ExecAllowlistMatcher.match(entries: [entry], resolution: resolution)
#expect(match?.pattern == entry.pattern)
}
@Test func matchUsesBasenameForSimplePattern() {
let entry = ExecAllowlistEntry(pattern: "rg")
let resolution = ExecCommandResolution(
rawExecutable: "rg",
resolvedPath: "/opt/homebrew/bin/rg",
executableName: "rg",
cwd: nil)
let match = ExecAllowlistMatcher.match(entries: [entry], resolution: resolution)
#expect(match?.pattern == entry.pattern)
}
@Test func matchIsCaseInsensitive() {
let entry = ExecAllowlistEntry(pattern: "RG")
let resolution = ExecCommandResolution(
rawExecutable: "rg",
resolvedPath: "/opt/homebrew/bin/rg",
executableName: "rg",
cwd: nil)
let match = ExecAllowlistMatcher.match(entries: [entry], resolution: resolution)
#expect(match?.pattern == entry.pattern)
}
@Test func matchSupportsGlobStar() {
let entry = ExecAllowlistEntry(pattern: "/opt/**/rg")
let resolution = ExecCommandResolution(
rawExecutable: "rg",
resolvedPath: "/opt/homebrew/bin/rg",
executableName: "rg",
cwd: nil)
let match = ExecAllowlistMatcher.match(entries: [entry], resolution: resolution)
#expect(match?.pattern == entry.pattern)
}
}

View File

@ -74,10 +74,6 @@ struct MacNodeRuntimeTests {
{
CLLocation(latitude: 0, longitude: 0)
}
func confirmSystemRun(command: String, cwd: String?) async -> SystemRunDecision {
.allowOnce
}
}
let services = await MainActor.run { FakeMainActorServices() }

View File

@ -0,0 +1,26 @@
import Foundation
import Testing
@testable import Clawdbot
@Suite(.serialized)
struct SessionMenuPreviewTests {
@Test func loaderReturnsCachedItems() async {
await SessionPreviewCache.shared._testReset()
let items = [SessionPreviewItem(id: "1", role: .user, text: "Hi")]
await SessionPreviewCache.shared._testSet(items: items, for: "main")
let snapshot = await SessionMenuPreviewLoader.load(sessionKey: "main", maxItems: 10)
#expect(snapshot.status == .ready)
#expect(snapshot.items.count == 1)
#expect(snapshot.items.first?.text == "Hi")
}
@Test func loaderReturnsEmptyWhenCachedEmpty() async {
await SessionPreviewCache.shared._testReset()
await SessionPreviewCache.shared._testSet(items: [], for: "main")
let snapshot = await SessionMenuPreviewLoader.load(sessionKey: "main", maxItems: 10)
#expect(snapshot.status == .empty)
#expect(snapshot.items.isEmpty)
}
}

View File

@ -37,7 +37,7 @@ import Testing
defaults.set(AppState.ConnectionMode.remote.rawValue, forKey: connectionModeKey)
defaults.set("ssh alice@example.com", forKey: remoteTargetKey)
let settings = CommandResolver.connectionSettings(defaults: defaults)
let settings = CommandResolver.connectionSettings(defaults: defaults, configRoot: [:])
#expect(settings.mode == .remote)
#expect(settings.target == "alice@example.com")
}

View File

@ -63,6 +63,8 @@ public struct BridgeHello: Codable, Sendable {
public let token: String?
public let platform: String?
public let version: String?
public let coreVersion: String?
public let uiVersion: String?
public let deviceFamily: String?
public let modelIdentifier: String?
public let caps: [String]?
@ -76,6 +78,8 @@ public struct BridgeHello: Codable, Sendable {
token: String?,
platform: String?,
version: String?,
coreVersion: String? = nil,
uiVersion: String? = nil,
deviceFamily: String? = nil,
modelIdentifier: String? = nil,
caps: [String]? = nil,
@ -88,6 +92,8 @@ public struct BridgeHello: Codable, Sendable {
self.token = token
self.platform = platform
self.version = version
self.coreVersion = coreVersion
self.uiVersion = uiVersion
self.deviceFamily = deviceFamily
self.modelIdentifier = modelIdentifier
self.caps = caps
@ -121,6 +127,8 @@ public struct BridgePairRequest: Codable, Sendable {
public let displayName: String?
public let platform: String?
public let version: String?
public let coreVersion: String?
public let uiVersion: String?
public let deviceFamily: String?
public let modelIdentifier: String?
public let caps: [String]?
@ -135,6 +143,8 @@ public struct BridgePairRequest: Codable, Sendable {
displayName: String?,
platform: String?,
version: String?,
coreVersion: String? = nil,
uiVersion: String? = nil,
deviceFamily: String? = nil,
modelIdentifier: String? = nil,
caps: [String]? = nil,
@ -148,6 +158,8 @@ public struct BridgePairRequest: Codable, Sendable {
self.displayName = displayName
self.platform = platform
self.version = version
self.coreVersion = coreVersion
self.uiVersion = uiVersion
self.deviceFamily = deviceFamily
self.modelIdentifier = modelIdentifier
self.caps = caps

View File

@ -4,6 +4,8 @@ public enum ClawdbotSystemCommand: String, Codable, Sendable {
case run = "system.run"
case which = "system.which"
case notify = "system.notify"
case execApprovalsGet = "system.execApprovals.get"
case execApprovalsSet = "system.execApprovals.set"
}
public enum ClawdbotNotificationPriority: String, Codable, Sendable {
@ -20,23 +22,32 @@ public enum ClawdbotNotificationDelivery: String, Codable, Sendable {
public struct ClawdbotSystemRunParams: Codable, Sendable, Equatable {
public var command: [String]
public var rawCommand: String?
public var cwd: String?
public var env: [String: String]?
public var timeoutMs: Int?
public var needsScreenRecording: Bool?
public var agentId: String?
public var sessionKey: String?
public init(
command: [String],
rawCommand: String? = nil,
cwd: String? = nil,
env: [String: String]? = nil,
timeoutMs: Int? = nil,
needsScreenRecording: Bool? = nil)
needsScreenRecording: Bool? = nil,
agentId: String? = nil,
sessionKey: String? = nil)
{
self.command = command
self.rawCommand = rawCommand
self.cwd = cwd
self.env = env
self.timeoutMs = timeoutMs
self.needsScreenRecording = needsScreenRecording
self.agentId = agentId
self.sessionKey = sessionKey
}
}

194
docs.acp.md Normal file
View File

@ -0,0 +1,194 @@
# Clawdbot ACP Bridge
This document describes how the Clawdbot ACP (Agent Client Protocol) bridge works,
how it maps ACP sessions to Gateway sessions, and how IDEs should invoke it.
## Overview
`clawdbot acp` exposes an ACP agent over stdio and forwards prompts to a running
Clawdbot Gateway over WebSocket. It keeps ACP session ids mapped to Gateway
session keys so IDEs can reconnect to the same agent transcript or reset it on
request.
Key goals:
- Minimal ACP surface area (stdio, NDJSON).
- Stable session mapping across reconnects.
- Works with existing Gateway session store (list/resolve/reset).
- Safe defaults (isolated ACP session keys by default).
## How can I use this
Use ACP when an IDE or tooling speaks Agent Client Protocol and you want it to
drive a Clawdbot Gateway session.
Quick steps:
1. Run a Gateway (local or remote).
2. Configure the Gateway target (`gateway.remote.url` + auth) or pass flags.
3. Point the IDE to run `clawdbot acp` over stdio.
Example config:
```bash
clawdbot config set gateway.remote.url wss://gateway-host:18789
clawdbot config set gateway.remote.token <token>
```
Example run:
```bash
clawdbot acp --url wss://gateway-host:18789 --token <token>
```
## Selecting agents
ACP does not pick agents directly. It routes by the Gateway session key.
Use agent-scoped session keys to target a specific agent:
```bash
clawdbot acp --session agent:main:main
clawdbot acp --session agent:design:main
clawdbot acp --session agent:qa:bug-123
```
Each ACP session maps to a single Gateway session key. One agent can have many
sessions; ACP defaults to an isolated `acp:<uuid>` session unless you override
the key or label.
## Zed editor setup
Add a custom ACP agent in `~/.config/zed/settings.json`:
```json
{
"agent_servers": {
"Clawdbot ACP": {
"type": "custom",
"command": "clawdbot",
"args": ["acp"],
"env": {}
}
}
}
```
To target a specific Gateway or agent:
```json
{
"agent_servers": {
"Clawdbot ACP": {
"type": "custom",
"command": "clawdbot",
"args": [
"acp",
"--url", "wss://gateway-host:18789",
"--token", "<token>",
"--session", "agent:design:main"
],
"env": {}
}
}
}
```
In Zed, open the Agent panel and select “Clawdbot ACP” to start a thread.
## Execution Model
- ACP client spawns `clawdbot acp` and speaks ACP messages over stdio.
- The bridge connects to the Gateway using existing auth config (or CLI flags).
- ACP `prompt` translates to Gateway `chat.send`.
- Gateway streaming events are translated back into ACP streaming events.
- ACP `cancel` maps to Gateway `chat.abort` for the active run.
## Session Mapping
By default each ACP session is mapped to a dedicated Gateway session key:
- `acp:<uuid>` unless overridden.
You can override or reuse sessions in two ways:
1) CLI defaults
```bash
clawdbot acp --session agent:main:main
clawdbot acp --session-label "support inbox"
clawdbot acp --reset-session
```
2) ACP metadata per session
```json
{
"_meta": {
"sessionKey": "agent:main:main",
"sessionLabel": "support inbox",
"resetSession": true,
"requireExisting": false
}
}
```
Rules:
- `sessionKey`: direct Gateway session key.
- `sessionLabel`: resolve an existing session by label.
- `resetSession`: mint a new transcript for the key before first use.
- `requireExisting`: fail if the key/label does not exist.
### Session Listing
ACP `listSessions` maps to Gateway `sessions.list` and returns a filtered
summary suitable for IDE session pickers. `_meta.limit` can cap the number of
sessions returned.
## Prompt Translation
ACP prompt inputs are converted into a Gateway `chat.send`:
- `text` and `resource` blocks become prompt text.
- `resource_link` with image mime types become attachments.
- The working directory can be prefixed into the prompt (default on, can be
disabled with `--no-prefix-cwd`).
Gateway streaming events are translated into ACP `message` and `tool_call`
updates. Terminal Gateway states map to ACP `done` with stop reasons:
- `complete` -> `stop`
- `aborted` -> `cancel`
- `error` -> `error`
## Auth + Gateway Discovery
`clawdbot acp` resolves the Gateway URL and auth from CLI flags or config:
- `--url` / `--token` / `--password` take precedence.
- Otherwise use configured `gateway.remote.*` settings.
## Operational Notes
- ACP sessions are stored in memory for the bridge process lifetime.
- Gateway session state is persisted by the Gateway itself.
- `--verbose` logs ACP/Gateway bridge events to stderr (never stdout).
- ACP runs can be canceled and the active run id is tracked per session.
## Compatibility
- ACP bridge uses `@agentclientprotocol/sdk` (currently 0.13.x).
- Works with ACP clients that implement `initialize`, `newSession`,
`loadSession`, `prompt`, `cancel`, and `listSessions`.
## Testing
- Unit: `src/acp/session.test.ts` covers run id lifecycle.
- Full gate: `pnpm lint && pnpm build && pnpm test && pnpm docs:build`.
## Related Docs
- CLI usage: `docs/cli/acp.md`
- Session model: `docs/concepts/session.md`
- Session management internals: `docs/reference/session-management-compaction.md`

40
docs/brave-search.md Normal file
View File

@ -0,0 +1,40 @@
---
summary: "Brave Search API setup for web_search"
read_when:
- You want to use Brave Search for web_search
- You need a BRAVE_API_KEY or plan details
---
# Brave Search API
Clawdbot uses Brave Search as the default provider for `web_search`.
## Get an API key
1) Create a Brave Search API account at https://brave.com/search/api/
2) In the dashboard, choose the **Data for Search** plan and generate an API key.
3) Store the key in config (recommended) or set `BRAVE_API_KEY` in the Gateway environment.
## Config example
```json5
{
tools: {
web: {
search: {
provider: "brave",
apiKey: "BRAVE_API_KEY_HERE",
maxResults: 5,
timeoutSeconds: 30
}
}
}
}
```
## Notes
- The Data for AI plan is **not** compatible with `web_search`.
- Brave provides a free tier plus paid plans; check the Brave API portal for current limits.
See [Web tools](/tools/web) for the full web_search configuration.

View File

@ -0,0 +1,64 @@
---
summary: "iMessage via BlueBubbles macOS server (REST send/receive, typing, reactions, pairing)."
read_when:
- Setting up BlueBubbles channel
- Troubleshooting webhook pairing
---
# BlueBubbles (macOS REST)
Status: bundled plugin (disabled by default) that talks to the BlueBubbles macOS server over HTTP.
## Overview
- Runs on macOS via the BlueBubbles helper app (`https://bluebubbles.app`).
- Clawdbot talks to it through its REST API (`GET /api/v1/ping`, `POST /message/text`, `POST /chat/:id/*`).
- Incoming messages arrive via webhooks; outgoing replies, typing indicators, read receipts, and tapbacks are REST calls.
- Attachments and stickers are ingested as inbound media (and surfaced to the agent when possible).
- Pairing/allowlist works the same way as other channels (`/start/pairing` etc) with `channels.bluebubbles.allowFrom` + pairing codes.
- Reactions are surfaced as system events just like Slack/Telegram so agents can “mention” them before replying.
## Quick start
1. Install the BlueBubbles server on your Mac (follows the app store instructions at `https://bluebubbles.app/install`).
2. In the BlueBubbles config, enable the web API and set a password for `guid`/`password`.
3. Configure Clawdbot:
```json5
{
channels: {
bluebubbles: {
enabled: true,
serverUrl: "http://bluebubbles-host:1234",
password: "example-password",
webhookPath: "/bluebubbles-webhook",
actions: { reactions: true }
}
}
}
```
4. Point BlueBubbles webhooks to your gateway (example: `http://your-gateway-host/bluebubbles-webhook?password=<password>`).
5. Start the gateway; it will register the webhook handler and start pairing.
## Configuration notes
- `channels.bluebubbles.serverUrl`: base URL of the BlueBubbles REST API.
- `channels.bluebubbles.password`: password that BlueBubbles expects on every request (`?password=...` or header).
- `channels.bluebubbles.webhookPath`: HTTP path the gateway exposes for BlueBubbles webhooks.
- `channels.bluebubbles.dmPolicy` / `groupPolicy` + `allowFrom`/`groupAllowFrom` behave like other channels; pairing/allowlist info is stored in `/pairing`.
- `channels.bluebubbles.actions.reactions` toggles whether the gateway enqueues system events for reactions/tapbacks.
- `channels.bluebubbles.textChunkLimit` overrides the default 4k limit.
- `channels.bluebubbles.mediaMaxMb` controls the max size of inbound attachments saved for analysis (default 8MB).
## How it works
- Outbound replies: `sendMessageBlueBubbles` resolves a chat GUID via `/api/v1/chat/query` and posts to `/api/v1/message/text`. Typing (`/api/v1/chat/<guid>/typing`) and read receipts (`/api/v1/chat/<guid>/read`) are sent before/after responses.
- Webhooks: BlueBubbles POSTs JSON payloads with `type` and `data`. The plugin ignores non-message events (typing indicator, read status) and extracts `chatGuid` from `data.chats[0].guid`.
- Reactions/tapbacks generate `BlueBubbles reaction added/removed` system events so agents can mention them. Agents can also trigger tapbacks via the `react` action with `messageId`, `emoji`, and a `to`/`chatGuid`.
- Attachments are downloaded via the REST API and stored in the inbound media cache; text-less messages are converted into `<media:...>` placeholders so the agent knows something was sent.
## Security
- Webhook requests are authenticated by comparing `guid`/`password` query params or headers against `channels.bluebubbles.password`. Requests from `localhost` are also accepted.
- Keep the API password and webhook endpoint secret (treat them like credentials).
- Enable HTTPS + firewall rules on the BlueBubbles server if exposing it outside your LAN.
## Troubleshooting
- If Voice/typing events stop working, check the BlueBubbles webhook logs and verify the gateway path matches `channels.bluebubbles.webhookPath`.
- Pairing codes expire after one hour; use `clawdbot pairing list bluebubbles` and `clawdbot pairing approve bluebubbles <code>`.
- Reactions require the BlueBubbles private API (`POST /api/v1/message/react`); ensure the server version exposes it.
For general channel workflow reference, see [/channels/index] and the [[plugins|/plugin]] guide.

View File

@ -58,7 +58,7 @@ Minimal config:
- The `discord` tool is only exposed when the current channel is Discord.
13. Native commands use isolated session keys (`agent:<agentId>:discord:slash:<userId>`) rather than the shared `main` session.
Note: Discord does not provide a simple username → id lookup without extra guild context, so prefer ids or `<@id>` mentions for DM delivery targets.
Note: Name → id resolution uses guild member search and requires Server Members Intent; if the bot cant search members, use ids or `<@id>` mentions.
Note: Slugs are lowercase with spaces replaced by `-`. Channel names are slugged without the leading `#`.
Note: Guild context `[from:]` lines include `author.tag` + `id` to make ping-ready replies easy.
@ -175,6 +175,7 @@ Notes:
- `agents.list[].groupChat.mentionPatterns` (or `messages.groupChat.mentionPatterns`) also count as mentions for guild messages.
- Multi-agent override: set per-agent patterns on `agents.list[].groupChat.mentionPatterns`.
- If `channels` is present, any channel not listed is denied by default.
- Threads inherit parent channel config (allowlist, `requireMention`, skills, prompts, etc.) unless you add the thread channel id explicitly.
- Bot-authored messages are ignored by default; set `channels.discord.allowBots=true` to allow them (own messages remain filtered).
- Warning: If you allow replies to other bots (`channels.discord.allowBots=true`), prevent bot-to-bot reply loops with `requireMention`, `channels.discord.guilds.*.channels.<id>.users` allowlists, and/or clear guardrails in `AGENTS.md` and `SOUL.md`.
@ -192,8 +193,11 @@ Notes:
- Your config requires mentions and you didnt mention it, or
- Your guild/channel allowlist denies the channel/user.
- **`requireMention: false` but still no replies**:
- `channels.discord.groupPolicy` defaults to **allowlist**; set it to `"open"` or add a guild entry under `channels.discord.guilds` (optionally list channels under `channels.discord.guilds.<id>.channels` to restrict).
- `requireMention` must live under `channels.discord.guilds` (or a specific channel). `channels.discord.requireMention` at the top level is ignored.
- `channels.discord.groupPolicy` defaults to **allowlist**; set it to `"open"` or add a guild entry under `channels.discord.guilds` (optionally list channels under `channels.discord.guilds.<id>.channels` to restrict).
- If you only set `DISCORD_BOT_TOKEN` and never create a `channels.discord` section, the runtime
defaults `groupPolicy` to `open`. Add `channels.discord.groupPolicy`,
`channels.defaults.groupPolicy`, or a guild/channel allowlist to lock it down.
- `requireMention` must live under `channels.discord.guilds` (or a specific channel). `channels.discord.requireMention` at the top level is ignored.
- **Permission audits** (`channels status --probe`) only check numeric channel IDs. If you use slugs/names as `channels.discord.guilds.*.channels` keys, the audit cant verify permissions.
- **DMs dont work**: `channels.discord.dm.enabled=false`, `channels.discord.dm.policy="disabled"`, or you havent been approved yet (`channels.discord.dm.policy="pairing"`).
@ -361,6 +365,10 @@ Allowlist matching notes:
- Use `*` to allow any sender/channel.
- When `guilds.<id>.channels` is present, channels not listed are denied by default.
- When `guilds.<id>.channels` is omitted, all channels in the allowlisted guild are allowed.
- To allow **no channels**, set `channels.discord.groupPolicy: "disabled"` (or keep an empty allowlist).
- The configure wizard accepts `Guild/Channel` names (public + private) and resolves them to IDs when possible.
- On startup, Clawdbot resolves channel/user names in allowlists to IDs (when the bot can search members)
and logs the mapping; unresolved entries are kept as typed.
Native command notes:
- The registered commands mirror Clawdbots chat commands.

View File

@ -17,6 +17,7 @@ Text is supported everywhere; media and reactions vary by channel.
- [Slack](/channels/slack) — Bolt SDK; workspace apps.
- [Signal](/channels/signal) — signal-cli; privacy-focused.
- [iMessage](/channels/imessage) — macOS only; native integration.
- [BlueBubbles](/channels/bluebubbles) — iMessage via BlueBubbles macOS server (bundled plugin, disabled by default).
- [Microsoft Teams](/channels/msteams) — Bot Framework; enterprise support (plugin, installed separately).
- [Matrix](/channels/matrix) — Matrix protocol (plugin, installed separately).
- [Zalo](/channels/zalo) — Zalo Bot API; Vietnam's popular messenger (plugin, installed separately).

View File

@ -70,9 +70,10 @@ Matrix is an open messaging protocol. Clawdbot connects as a Matrix user and lis
- `clawdbot pairing list matrix`
- `clawdbot pairing approve matrix <CODE>`
- Public DMs: `channels.matrix.dm.policy="open"` plus `channels.matrix.dm.allowFrom=["*"]`.
- `channels.matrix.dm.allowFrom` accepts user IDs or display names (resolved at startup when directory search is available).
## Rooms (groups)
- Default: `channels.matrix.groupPolicy = "allowlist"` (mention-gated).
- Default: `channels.matrix.groupPolicy = "allowlist"` (mention-gated). Use `channels.defaults.groupPolicy` to override the default when unset.
- Allowlist rooms with `channels.matrix.rooms`:
```json5
{
@ -86,6 +87,9 @@ Matrix is an open messaging protocol. Clawdbot connects as a Matrix user and lis
}
```
- `requireMention: false` enables auto-reply in that room.
- The configure wizard prompts for room allowlists (room IDs, aliases, or names) and resolves names when possible.
- On startup, Clawdbot resolves room/user names in allowlists to IDs and logs the mapping; unresolved entries are kept as typed.
- To allow **no rooms**, set `channels.matrix.groupPolicy: "disabled"` (or keep an empty allowlist).
## Threads
- Reply threading is supported.

View File

@ -76,12 +76,13 @@ Disable with:
**DM access**
- Default: `channels.msteams.dmPolicy = "pairing"`. Unknown senders are ignored until approved.
- `channels.msteams.allowFrom` accepts AAD object IDs or UPNs.
- `channels.msteams.allowFrom` accepts AAD object IDs, UPNs, or display names (resolved at startup when Graph allows).
**Group access**
- Default: `channels.msteams.groupPolicy = "allowlist"` (blocked unless you add `groupAllowFrom`).
- Default: `channels.msteams.groupPolicy = "allowlist"` (blocked unless you add `groupAllowFrom`). Use `channels.defaults.groupPolicy` to override the default when unset.
- `channels.msteams.groupAllowFrom` controls which senders can trigger in group chats/channels (falls back to `channels.msteams.allowFrom`).
- Set `groupPolicy: "open"` to allow any member (still mentiongated by default).
- To allow **no channels**, set `channels.msteams.groupPolicy: "disabled"`.
Example:
```json5
@ -95,6 +96,32 @@ Example:
}
```
**Teams + channel allowlist**
- Scope group/channel replies by listing teams and channels under `channels.msteams.teams`.
- Keys can be team IDs or names; channel keys can be conversation IDs or names.
- When `groupPolicy="allowlist"` and a teams allowlist is present, only listed teams/channels are accepted (mentiongated).
- The configure wizard accepts `Team/Channel` entries and stores them for you.
- On startup, Clawdbot resolves team/channel and user allowlist names to IDs (when Graph permissions allow)
and logs the mapping; unresolved entries are kept as typed.
Example:
```json5
{
channels: {
msteams: {
groupPolicy: "allowlist",
teams: {
"My Team": {
channels: {
"General": { requireMention: true }
}
}
}
}
}
}
```
## How it works
1. Install the Microsoft Teams plugin.
2. Create an **Azure Bot** (App ID + secret + tenant ID).

View File

@ -1,11 +1,13 @@
---
summary: "Slack socket mode setup and Clawdbot config"
read_when: "Setting up Slack or debugging Slack socket mode"
summary: "Slack setup for socket or HTTP webhook mode"
read_when: "Setting up Slack or debugging Slack socket/HTTP mode"
---
# Slack (socket mode)
# Slack
## Quick setup (beginner)
## Socket mode (default)
### Quick setup (beginner)
1) Create a Slack app and enable **Socket Mode**.
2) Create an **App Token** (`xapp-...`) and **Bot Token** (`xoxb-...`).
3) Set tokens for Clawdbot and start the gateway.
@ -23,7 +25,7 @@ Minimal config:
}
```
## Setup
### Setup
1) Create a Slack app (From scratch) in https://api.channels.slack.com/apps.
2) **Socket Mode** → toggle on. Then go to **Basic Information****App-Level Tokens****Generate Token and Scopes** with scope `connections:write`. Copy the **App Token** (`xapp-...`).
3) **OAuth & Permissions** → add bot token scopes (use the manifest below). Click **Install to Workspace**. Copy the **Bot User OAuth Token** (`xoxb-...`).
@ -43,7 +45,7 @@ Use the manifest below so scopes and events stay in sync.
Multi-account support: use `channels.slack.accounts` with per-account tokens and optional `name`. See [`gateway/configuration`](/gateway/configuration#telegramaccounts--discordaccounts--slackaccounts--signalaccounts--imessageaccounts) for the shared pattern.
## Clawdbot config (minimal)
### Clawdbot config (minimal)
Set tokens via env vars (recommended):
- `SLACK_APP_TOKEN=xapp-...`
@ -63,7 +65,7 @@ Or via config:
}
```
## User token (optional)
### User token (optional)
Clawdbot can use a Slack user token (`xoxp-...`) for read operations (history,
pins, reactions, emoji, member info). By default this stays read-only: reads
prefer the user token when present, and writes still use the bot token unless
@ -102,18 +104,51 @@ Example with userTokenReadOnly explicitly set (allow user token writes):
}
```
### Token usage
#### Token usage
- Read operations (history, reactions list, pins list, emoji list, member info,
search) prefer the user token when configured, otherwise the bot token.
- Write operations (send/edit/delete messages, add/remove reactions, pin/unpin,
file uploads) use the bot token by default. If `userTokenReadOnly: false` and
no bot token is available, Clawdbot falls back to the user token.
## History context
### History context
- `channels.slack.historyLimit` (or `channels.slack.accounts.*.historyLimit`) controls how many recent channel/group messages are wrapped into the prompt.
- Falls back to `messages.groupChat.historyLimit`. Set `0` to disable (default 50).
## Manifest (optional)
## HTTP mode (Events API)
Use HTTP webhook mode when your Gateway is reachable by Slack over HTTPS (typical for server deployments).
HTTP mode uses the Events API + Interactivity + Slash Commands with a shared request URL.
### Setup
1) Create a Slack app and **disable Socket Mode** (optional if you only use HTTP).
2) **Basic Information** → copy the **Signing Secret**.
3) **OAuth & Permissions** → install the app and copy the **Bot User OAuth Token** (`xoxb-...`).
4) **Event Subscriptions** → enable events and set the **Request URL** to your gateway webhook path (default `/slack/events`).
5) **Interactivity & Shortcuts** → enable and set the same **Request URL**.
6) **Slash Commands** → set the same **Request URL** for your command(s).
Example request URL:
`https://gateway-host/slack/events`
### Clawdbot config (minimal)
```json5
{
channels: {
slack: {
enabled: true,
mode: "http",
botToken: "xoxb-...",
signingSecret: "your-signing-secret",
webhookPath: "/slack/events"
}
}
}
```
Multi-account HTTP mode: set `channels.slack.accounts.<id>.mode = "http"` and provide a unique
`webhookPath` per account so each Slack app can point to its own URL.
### Manifest (optional)
Use this Slack app manifest to create the app quickly (adjust the name/command if you want). Include the
user scopes if you plan to configure a user token.
@ -343,10 +378,19 @@ For fine-grained control, use these tags in agent responses:
- Default: `channels.slack.dm.policy="pairing"` — unknown DM senders get a pairing code (expires after 1 hour).
- Approve via: `clawdbot pairing approve slack <code>`.
- To allow anyone: set `channels.slack.dm.policy="open"` and `channels.slack.dm.allowFrom=["*"]`.
- `channels.slack.dm.allowFrom` accepts user IDs, @handles, or emails (resolved at startup when tokens allow).
## Group policy
- `channels.slack.groupPolicy` controls channel handling (`open|disabled|allowlist`).
- `allowlist` requires channels to be listed in `channels.slack.channels`.
- If you only set `SLACK_BOT_TOKEN`/`SLACK_APP_TOKEN` and never create a `channels.slack` section,
the runtime defaults `groupPolicy` to `open`. Add `channels.slack.groupPolicy`,
`channels.defaults.groupPolicy`, or a channel allowlist to lock it down.
- The configure wizard accepts `#channel` names and resolves them to IDs when possible
(public + private); if multiple matches exist, it prefers the active channel.
- On startup, Clawdbot resolves channel/user names in allowlists to IDs (when tokens allow)
and logs the mapping; unresolved entries are kept as typed.
- To allow **no channels**, set `channels.slack.groupPolicy: "disabled"` (or keep an empty allowlist).
Channel options (`channels.slack.channels.<id>` or `channels.slack.channels.<name>`):
- `allow`: allow/deny the channel when `groupPolicy="allowlist"`.

View File

@ -152,6 +152,7 @@ By default, the bot only responds to mentions in groups (`@botname` or patterns
```
**Important:** Setting `channels.telegram.groups` creates an **allowlist** - only listed groups (or `"*"`) will be accepted.
Forum topics inherit their parent group config (allowFrom, requireMention, skills, prompts) unless you add per-topic overrides under `channels.telegram.groups.<groupId>.topics.<topicId>`.
To allow all groups with always-respond:
```json5
@ -216,6 +217,7 @@ Telegram forum topics include a `message_thread_id` per message. Clawdbot:
- General topic (thread id `1`) is special: message sends omit `message_thread_id` (Telegram rejects it), but typing indicators still include it.
- Exposes `MessageThreadId` + `IsForum` in template context for routing/templating.
- Topic-specific configuration is available under `channels.telegram.groups.<chatId>.topics.<threadId>` (skills, allowlists, auto-reply, system prompts, disable).
- Topic configs inherit group settings (requireMention, allowlists, skills, prompts, enabled) unless overridden per topic.
Private chats can include `message_thread_id` in some edge cases. Clawdbot keeps the DM session key unchanged, but still uses the thread id for replies/draft streaming when it is present.

View File

@ -66,11 +66,36 @@ clawdbot directory groups list --channel zalouser --query "work"
## Access control (DMs)
`channels.zalouser.dmPolicy` supports: `pairing | allowlist | open | disabled` (default: `pairing`).
`channels.zalouser.allowFrom` accepts user IDs or names (resolved at startup when available).
Approve via:
- `clawdbot pairing list zalouser`
- `clawdbot pairing approve zalouser <code>`
## Group access (optional)
- Default: `channels.zalouser.groupPolicy = "open"` (groups allowed). Use `channels.defaults.groupPolicy` to override the default when unset.
- Restrict to an allowlist with:
- `channels.zalouser.groupPolicy = "allowlist"`
- `channels.zalouser.groups` (keys are group IDs or names)
- Block all groups: `channels.zalouser.groupPolicy = "disabled"`.
- The configure wizard can prompt for group allowlists.
- On startup, Clawdbot resolves group/user names in allowlists to IDs and logs the mapping; unresolved entries are kept as typed.
Example:
```json5
{
channels: {
zalouser: {
groupPolicy: "allowlist",
groups: {
"123456789": { allow: true },
"Work Chat": { allow: true }
}
}
}
}
```
## Multi-account
Accounts map to zca profiles. Example:

166
docs/cli/acp.md Normal file
View File

@ -0,0 +1,166 @@
---
summary: "Run the ACP bridge for IDE integrations"
read_when:
- Setting up ACP-based IDE integrations
- Debugging ACP session routing to the Gateway
---
# acp
Run the ACP (Agent Client Protocol) bridge that talks to a Clawdbot Gateway.
This command speaks ACP over stdio for IDEs and forwards prompts to the Gateway
over WebSocket. It keeps ACP sessions mapped to Gateway session keys.
## Usage
```bash
clawdbot acp
# Remote Gateway
clawdbot acp --url wss://gateway-host:18789 --token <token>
# Attach to an existing session key
clawdbot acp --session agent:main:main
# Attach by label (must already exist)
clawdbot acp --session-label "support inbox"
# Reset the session key before the first prompt
clawdbot acp --session agent:main:main --reset-session
```
## ACP client (debug)
Use the built-in ACP client to sanity-check the bridge without an IDE.
It spawns the ACP bridge and lets you type prompts interactively.
```bash
clawdbot acp client
# Point the spawned bridge at a remote Gateway
clawdbot acp client --server-args --url wss://gateway-host:18789 --token <token>
# Override the server command (default: clawdbot)
clawdbot acp client --server "node" --server-args dist/entry.js acp --url ws://127.0.0.1:19001
```
## How to use this
Use ACP when an IDE (or other client) speaks Agent Client Protocol and you want
it to drive a Clawdbot Gateway session.
1. Ensure the Gateway is running (local or remote).
2. Configure the Gateway target (config or flags).
3. Point your IDE to run `clawdbot acp` over stdio.
Example config (persisted):
```bash
clawdbot config set gateway.remote.url wss://gateway-host:18789
clawdbot config set gateway.remote.token <token>
```
Example direct run (no config write):
```bash
clawdbot acp --url wss://gateway-host:18789 --token <token>
```
## Selecting agents
ACP does not pick agents directly. It routes by the Gateway session key.
Use agent-scoped session keys to target a specific agent:
```bash
clawdbot acp --session agent:main:main
clawdbot acp --session agent:design:main
clawdbot acp --session agent:qa:bug-123
```
Each ACP session maps to a single Gateway session key. One agent can have many
sessions; ACP defaults to an isolated `acp:<uuid>` session unless you override
the key or label.
## Zed editor setup
Add a custom ACP agent in `~/.config/zed/settings.json` (or use Zeds Settings UI):
```json
{
"agent_servers": {
"Clawdbot ACP": {
"type": "custom",
"command": "clawdbot",
"args": ["acp"],
"env": {}
}
}
}
```
To target a specific Gateway or agent:
```json
{
"agent_servers": {
"Clawdbot ACP": {
"type": "custom",
"command": "clawdbot",
"args": [
"acp",
"--url", "wss://gateway-host:18789",
"--token", "<token>",
"--session", "agent:design:main"
],
"env": {}
}
}
}
```
In Zed, open the Agent panel and select “Clawdbot ACP” to start a thread.
## Session mapping
By default, ACP sessions get an isolated Gateway session key with an `acp:` prefix.
To reuse a known session, pass a session key or label:
- `--session <key>`: use a specific Gateway session key.
- `--session-label <label>`: resolve an existing session by label.
- `--reset-session`: mint a fresh session id for that key (same key, new transcript).
If your ACP client supports metadata, you can override per session:
```json
{
"_meta": {
"sessionKey": "agent:main:main",
"sessionLabel": "support inbox",
"resetSession": true
}
}
```
Learn more about session keys at [/concepts/session](/concepts/session).
## Options
- `--url <url>`: Gateway WebSocket URL (defaults to gateway.remote.url when configured).
- `--token <token>`: Gateway auth token.
- `--password <password>`: Gateway auth password.
- `--session <key>`: default session key.
- `--session-label <label>`: default session label to resolve.
- `--require-existing`: fail if the session key/label does not exist.
- `--reset-session`: reset the session key before first use.
- `--no-prefix-cwd`: do not prefix prompts with the working directory.
- `--verbose, -v`: verbose logging to stderr.
### `acp client` options
- `--cwd <dir>`: working directory for the ACP session.
- `--server <command>`: ACP server command (default: `clawdbot`).
- `--server-args <args...>`: extra arguments passed to the ACP server.
- `--server-verbose`: enable verbose logging on the ACP server.
- `--verbose, -v`: verbose client logging.

44
docs/cli/approvals.md Normal file
View File

@ -0,0 +1,44 @@
---
summary: "CLI reference for `clawdbot approvals` (exec approvals for gateway or node hosts)"
read_when:
- You want to edit exec approvals from the CLI
- You need to manage allowlists on gateway or node hosts
---
# `clawdbot approvals`
Manage exec approvals for the **gateway host** or a **node host**.
By default, commands target the gateway. Use `--node` to edit a nodes approvals.
Related:
- Exec approvals: [Exec approvals](/tools/exec-approvals)
- Nodes: [Nodes](/nodes)
## Common commands
```bash
clawdbot approvals get
clawdbot approvals get --node <id|name|ip>
```
## Replace approvals from a file
```bash
clawdbot approvals set --file ./exec-approvals.json
clawdbot approvals set --node <id|name|ip> --file ./exec-approvals.json
```
## Allowlist helpers
```bash
clawdbot approvals allowlist add "~/Projects/**/bin/rg"
clawdbot approvals allowlist add --agent main --node <id|name|ip> "/usr/bin/uptime"
clawdbot approvals allowlist remove "~/Projects/**/bin/rg"
```
## Notes
- `--node` uses the same resolver as `clawdbot nodes` (id, name, ip, or id prefix).
- The node host must advertise `system.execApprovals.get/set` (macOS app or headless node host).
- Approvals files are stored per host at `~/.clawdbot/exec-approvals.json`.

View File

@ -20,6 +20,7 @@ clawdbot channels list
clawdbot channels status
clawdbot channels capabilities
clawdbot channels capabilities --channel discord --target channel:123
clawdbot channels resolve --channel slack "#general" "@jane"
clawdbot channels logs --channel all
```
@ -57,3 +58,17 @@ Notes:
- `--channel` is optional; omit it to list every channel (including extensions).
- `--target` accepts `channel:<id>` or a raw numeric channel id and only applies to Discord.
- Probes are provider-specific: Discord intents + optional channel permissions; Slack bot + user scopes; Telegram bot flags + webhook; Signal daemon version; MS Teams app token + Graph roles/scopes (annotated where known). Channels without probes report `Probe: unavailable`.
## Resolve names to IDs
Resolve channel/user names to IDs using the provider directory:
```bash
clawdbot channels resolve --channel slack "#general" "@jane"
clawdbot channels resolve --channel discord "My Server/#support" "@someone"
clawdbot channels resolve --channel matrix "Project Room"
```
Notes:
- Use `--kind user|group|auto` to force the target type.
- Resolution prefers active matches when multiple entries share the same name.

View File

@ -15,6 +15,7 @@ the configure wizard (same as `clawdbot configure`).
clawdbot config get browser.executablePath
clawdbot config set browser.executablePath "/usr/bin/google-chrome"
clawdbot config set agents.defaults.heartbeat.every "2h"
clawdbot config set agents.list[0].tools.exec.node "node-id-or-name"
clawdbot config unset tools.web.search.apiKey
```
@ -27,6 +28,13 @@ clawdbot config get agents.defaults.workspace
clawdbot config get agents.list[0].id
```
Use the agent list index to target a specific agent:
```bash
clawdbot config get agents.list
clawdbot config set agents.list[1].tools.exec.node "node-id-or-name"
```
## Values
Values are parsed as JSON5 when possible; otherwise they are treated as strings.

View File

@ -17,6 +17,7 @@ Related:
Notes:
- Choosing where the Gateway runs always updates `gateway.mode`. You can select "Continue" without other sections if that is all you need.
- Channel-oriented services (Slack/Discord/Matrix/Microsoft Teams) prompt for channel/room allowlists during setup. You can enter names or IDs; the wizard resolves names to IDs when possible.
## Examples

View File

@ -9,6 +9,9 @@ read_when:
Manage the Gateway daemon (background service).
Note: `clawdbot service gateway …` is the preferred surface; `daemon` remains
as a legacy alias for compatibility.
Related:
- Gateway CLI: [Gateway](/cli/gateway)
- macOS platform notes: [macOS](/platforms/macos)

View File

@ -7,10 +7,11 @@ read_when:
# `clawdbot hooks`
Manage agent hooks (event-driven automations for commands like `/new`, `/reset`, etc.).
Manage agent hooks (event-driven automations for commands like `/new`, `/reset`, and gateway startup).
Related:
- Hooks: [Hooks](/hooks)
- Plugin hooks: [Plugins](/plugin#plugin-hooks)
## List All Hooks
@ -28,11 +29,13 @@ List all discovered hooks from workspace, managed, and bundled directories.
**Example output:**
```
Hooks (2/2 ready)
Hooks (4/4 ready)
Ready:
🚀 boot-md ✓ - Run BOOT.md on gateway startup
📝 command-logger ✓ - Log all command events to a centralized audit file
💾 session-memory ✓ - Save session context to memory when /new command is issued
😈 soul-evil ✓ - Swap injected SOUL content during a purge window or by random chance
```
**Example (verbose):**
@ -105,8 +108,8 @@ Show summary of hook eligibility status (how many are ready vs. not ready).
```
Hooks Status
Total hooks: 2
Ready: 2
Total hooks: 4
Ready: 4
Not ready: 0
```
@ -118,6 +121,9 @@ clawdbot hooks enable <name>
Enable a specific hook by adding it to your config (`~/.clawdbot/config.json`).
**Note:** Hooks managed by plugins show `plugin:<id>` in `clawdbot hooks list` and
cant be enabled/disabled here. Enable/disable the plugin instead.
**Arguments:**
- `<name>`: Hook name (e.g., `session-memory`)
@ -256,3 +262,29 @@ grep '"action":"new"' ~/.clawdbot/logs/commands.log | jq .
```
**See:** [command-logger documentation](/hooks#command-logger)
### soul-evil
Swaps injected `SOUL.md` content with `SOUL_EVIL.md` during a purge window or by random chance.
**Enable:**
```bash
clawdbot hooks enable soul-evil
```
**See:** [SOUL Evil Hook](/hooks/soul-evil)
### boot-md
Runs `BOOT.md` when the gateway starts (after channels start).
**Events**: `gateway:startup`
**Enable**:
```bash
clawdbot hooks enable boot-md
```
**See:** [boot-md documentation](/hooks#boot-md)

View File

@ -23,15 +23,19 @@ This page describes the current CLI behavior. If commands change, update this do
- [`message`](/cli/message)
- [`agent`](/cli/agent)
- [`agents`](/cli/agents)
- [`acp`](/cli/acp)
- [`status`](/cli/status)
- [`health`](/cli/health)
- [`sessions`](/cli/sessions)
- [`gateway`](/cli/gateway)
- [`daemon`](/cli/daemon)
- [`service`](/cli/service)
- [`logs`](/cli/logs)
- [`models`](/cli/models)
- [`memory`](/cli/memory)
- [`nodes`](/cli/nodes)
- [`node`](/cli/node)
- [`approvals`](/cli/approvals)
- [`sandbox`](/cli/sandbox)
- [`tui`](/cli/tui)
- [`browser`](/cli/browser)
@ -125,6 +129,7 @@ clawdbot [--dev] [--profile <name>] <command>
list
add
delete
acp
status
health
sessions
@ -140,6 +145,21 @@ clawdbot [--dev] [--profile <name>] <command>
start
stop
restart
service
gateway
status
install
uninstall
start
stop
restart
node
status
install
uninstall
start
stop
restart
logs
models
list
@ -168,21 +188,19 @@ clawdbot [--dev] [--profile <name>] <command>
runs
run
nodes
status
describe
list
pending
approve
reject
rename
invoke
run
notify
camera list|snap|clip
canvas snapshot|present|hide|navigate|eval
canvas a2ui push|reset
screen record
location get
node
start
daemon
status
install
uninstall
start
stop
restart
approvals
get
set
allowlist add|remove
browser
status
start
@ -506,6 +524,11 @@ Options:
- `--force`
- `--json`
### `acp`
Run the ACP bridge that connects IDEs to the Gateway.
See [`acp`](/cli/acp) for full options and examples.
### `status`
Show linked session health and recent recipients.
@ -518,11 +541,14 @@ Options:
- `--verbose`
- `--debug` (alias for `--verbose`)
Notes:
- Overview includes Gateway + Node service status when available.
### Usage tracking
Clawdbot can surface provider usage/quota when OAuth/API creds are available.
Surfaces:
- `/status` (alias: `/usage`; adds a short usage line when available)
- `/status` (adds a short provider usage line when available)
- `clawdbot status --usage` (prints full provider breakdown)
- macOS menu bar (Usage section under Context)
@ -772,6 +798,23 @@ Subcommands:
All `cron` commands accept `--url`, `--token`, `--timeout`, `--expect-final`.
## Node host
`node` runs a **headless node host** or manages it as a background service. See
[`clawdbot node`](/cli/node).
Subcommands:
- `node start --host <gateway-host> --port 18790`
- `node service status`
- `node service install [--host <gateway-host>] [--port <port>] [--tls] [--tls-fingerprint <sha256>] [--node-id <id>] [--display-name <name>] [--runtime <node|bun>] [--force]`
- `node service uninstall`
- `node service start`
- `node service stop`
- `node service restart`
Legacy alias:
- `node daemon …` (same as `node service …`)
## Nodes
`nodes` talks to the Gateway and targets paired nodes. See [/nodes](/nodes).
@ -788,7 +831,7 @@ Subcommands:
- `nodes reject <requestId>`
- `nodes rename --node <id|name|ip> --name <displayName>`
- `nodes invoke --node <id|name|ip> --command <command> [--params <json>] [--invoke-timeout <ms>] [--idempotency-key <key>]`
- `nodes run --node <id|name|ip> [--cwd <path>] [--env KEY=VAL] [--command-timeout <ms>] [--needs-screen-recording] [--invoke-timeout <ms>] <command...>` (mac only)
- `nodes run --node <id|name|ip> [--cwd <path>] [--env KEY=VAL] [--command-timeout <ms>] [--needs-screen-recording] [--invoke-timeout <ms>] <command...>` (mac node or headless node host)
- `nodes notify --node <id|name|ip> [--title <text>] [--body <text>] [--sound <name>] [--priority <passive|active|timeSensitive>] [--delivery <system|overlay|auto>] [--invoke-timeout <ms>]` (mac only)
Camera:

View File

@ -7,10 +7,12 @@ read_when:
# `clawdbot memory`
Memory search tools (semantic memory status/index/search).
Manage semantic memory indexing and search.
Provided by the active memory plugin (default: `memory-core`; set `plugins.slots.memory = "none"` to disable).
Related:
- Memory concept: [Memory](/concepts/memory)
- Plugins: [Plugins](/plugins)
## Examples
@ -18,6 +20,22 @@ Related:
clawdbot memory status
clawdbot memory status --deep
clawdbot memory status --deep --index
clawdbot memory status --deep --index --verbose
clawdbot memory index
clawdbot memory index --verbose
clawdbot memory search "release checklist"
clawdbot memory status --agent main
clawdbot memory index --agent main --verbose
```
## Options
Common:
- `--agent <id>`: scope to a single agent (default: all configured agents).
- `--verbose`: emit detailed logs during probes and indexing.
Notes:
- `memory status --deep` probes vector + embedding availability.
- `memory status --deep --index` runs a reindex if the store is dirty.
- `memory index --verbose` prints per-phase details (provider, model, sources, batch activity).

96
docs/cli/node.md Normal file
View File

@ -0,0 +1,96 @@
---
summary: "CLI reference for `clawdbot node` (headless node host)"
read_when:
- Running the headless node host
- Pairing a non-macOS node for system.run
---
# `clawdbot node`
Run a **headless node host** that connects to the Gateway bridge and exposes
`system.run` / `system.which` on this machine.
## Why use a node host?
Use a node host when you want agents to **run commands on other machines** in your
network without installing a full macOS companion app there.
Common use cases:
- Run commands on remote Linux/Windows boxes (build servers, lab machines, NAS).
- Keep exec **sandboxed** on the gateway, but delegate approved runs to other hosts.
- Provide a lightweight, headless execution target for automation or CI nodes.
Execution is still guarded by **exec approvals** and peragent allowlists on the
node host, so you can keep command access scoped and explicit.
## Start (foreground)
```bash
clawdbot node start --host <gateway-host> --port 18790
```
Options:
- `--host <host>`: Gateway bridge host (default: `127.0.0.1`)
- `--port <port>`: Gateway bridge port (default: `18790`)
- `--tls`: Use TLS for the bridge connection
- `--tls-fingerprint <sha256>`: Pin the bridge certificate fingerprint
- `--node-id <id>`: Override node id (clears pairing token)
- `--display-name <name>`: Override the node display name
## Service (background)
Install a headless node host as a user service.
```bash
clawdbot node service install --host <gateway-host> --port 18790
# or
clawdbot service node install --host <gateway-host> --port 18790
```
Options:
- `--host <host>`: Gateway bridge host (default: `127.0.0.1`)
- `--port <port>`: Gateway bridge port (default: `18790`)
- `--tls`: Use TLS for the bridge connection
- `--tls-fingerprint <sha256>`: Pin the bridge certificate fingerprint
- `--node-id <id>`: Override node id (clears pairing token)
- `--display-name <name>`: Override the node display name
- `--runtime <runtime>`: Service runtime (`node` or `bun`)
- `--force`: Reinstall/overwrite if already installed
Manage the service:
```bash
clawdbot node status
clawdbot service node status
clawdbot node service status
clawdbot node service start
clawdbot node service stop
clawdbot node service restart
clawdbot node service uninstall
```
Legacy alias:
```bash
clawdbot node daemon status
```
## Pairing
The first connection creates a pending node pair request on the Gateway.
Approve it via:
```bash
clawdbot nodes pending
clawdbot nodes approve <requestId>
```
The node host stores its node id + token in `~/.clawdbot/node.json`.
## Exec approvals
`system.run` is gated by local exec approvals:
- `~/.clawdbot/exec-approvals.json`
- [Exec approvals](/tools/exec-approvals)
- `clawdbot approvals --node <id|name|ip>` (edit from the Gateway)

51
docs/cli/service.md Normal file
View File

@ -0,0 +1,51 @@
---
summary: "CLI reference for `clawdbot service` (manage gateway + node services)"
read_when:
- You want to manage Gateway or node services cross-platform
- You want a single surface for start/stop/install/uninstall
---
# `clawdbot service`
Manage the **Gateway** service and **node host** services.
Related:
- Gateway daemon (legacy alias): [Daemon](/cli/daemon)
- Node host: [Node](/cli/node)
## Gateway service
```bash
clawdbot service gateway status
clawdbot service gateway install --port 18789
clawdbot service gateway start
clawdbot service gateway stop
clawdbot service gateway restart
clawdbot service gateway uninstall
```
Notes:
- `service gateway status` supports `--json` and `--deep` for system checks.
- `service gateway install` supports `--runtime node|bun` and `--token`.
## Node host service
```bash
clawdbot service node status
clawdbot service node install --host <gateway-host> --port 18790
clawdbot service node start
clawdbot service node stop
clawdbot service node restart
clawdbot service node uninstall
```
Notes:
- `service node install` supports `--runtime node|bun`, `--node-id`, `--display-name`,
and TLS options (`--tls`, `--tls-fingerprint`).
## Aliases
- `clawdbot daemon …``clawdbot service gateway …`
- `clawdbot node service …``clawdbot service node …`
- `clawdbot node status``clawdbot service node status`
- `clawdbot node daemon …``clawdbot service node …` (legacy)

View File

@ -19,4 +19,5 @@ clawdbot status --usage
Notes:
- `--deep` runs live probes (WhatsApp Web + Telegram + Discord + Slack + Signal).
- Output includes per-agent session stores when multiple agents are configured.
- Overview includes Gateway + Node service install/runtime status when available.
- Update info surfaces in the Overview; if an update is available, status prints a hint to run `clawdbot update` (see [Updating](/install/updating)).

View File

@ -5,13 +5,19 @@ read_when:
---
# Agent Loop (Clawdbot)
Short, exact flow of one agent run.
An agentic loop is the full “real” run of an agent: intake → context assembly → model inference →
tool execution → streaming replies → persistence. Its the authoritative path that turns a message
into actions and a final reply, while keeping session state consistent.
In Clawdbot, a loop is a single, serialized run per session that emits lifecycle and stream events
as the model thinks, calls tools, and streams output. This doc explains how that authentic loop is
wired end-to-end.
## Entry points
- Gateway RPC: `agent` and `agent.wait`.
- CLI: `agent` command.
## High-level flow
## How it works (high-level)
1) `agent` RPC validates params, resolves session (sessionKey/sessionId), persists session metadata, returns `{ runId, acceptedAt }` immediately.
2) `agentCommand` runs the agent:
- resolves model + thinking/verbose defaults
@ -19,8 +25,9 @@ Short, exact flow of one agent run.
- calls `runEmbeddedPiAgent` (pi-agent-core runtime)
- emits **lifecycle end/error** if the embedded loop does not emit one
3) `runEmbeddedPiAgent`:
- builds `AgentSession` and subscribes to pi events
- streams assistant deltas + tool events
- serializes runs via per-session + global queues
- resolves model + auth profile and builds the pi session
- subscribes to pi events and streams assistant/tool deltas
- enforces timeout -> aborts run if exceeded
- returns payloads + usage metadata
4) `subscribeEmbeddedPiSession` bridges pi-agent-core events to Clawdbot `agent` stream:
@ -31,6 +38,73 @@ Short, exact flow of one agent run.
- waits for **lifecycle end/error** for `runId`
- returns `{ status: ok|error|timeout, startedAt, endedAt, error? }`
## Queueing + concurrency
- Runs are serialized per session key (session lane) and optionally through a global lane.
- This prevents tool/session races and keeps session history consistent.
- Messaging channels can choose queue modes (collect/steer/followup) that feed this lane system.
See [Command Queue](/concepts/queue).
## Session + workspace preparation
- Workspace is resolved and created; sandboxed runs may redirect to a sandbox workspace root.
- Skills are loaded (or reused from a snapshot) and injected into env and prompt.
- Bootstrap/context files are resolved and injected into the system prompt report.
- A session write lock is acquired; `SessionManager` is opened and prepared before streaming.
## Prompt assembly + system prompt
- System prompt is built from Clawdbots base prompt, skills prompt, bootstrap context, and per-run overrides.
- Model-specific limits and compaction reserve tokens are enforced.
- See [System prompt](/concepts/system-prompt) for what the model sees.
## Hook points (where you can intercept)
Clawdbot has two hook systems:
- **Internal hooks** (Gateway hooks): event-driven scripts for commands and lifecycle events.
- **Plugin hooks**: extension points inside the agent/tool lifecycle and gateway pipeline.
### Internal hooks (Gateway hooks)
- **`agent:bootstrap`**: runs while building bootstrap files before the system prompt is finalized.
Use this to add/remove bootstrap context files.
- **Command hooks**: `/new`, `/reset`, `/stop`, and other command events (see Hooks doc).
See [Hooks](/hooks) for setup and examples.
### Plugin hooks (agent + gateway lifecycle)
These run inside the agent loop or gateway pipeline:
- **`before_agent_start`**: inject context or override system prompt before the run starts.
- **`agent_end`**: inspect the final message list and run metadata after completion.
- **`before_compaction` / `after_compaction`**: observe or annotate compaction cycles.
- **`before_tool_call` / `after_tool_call`**: intercept tool params/results.
- **`message_received` / `message_sending` / `message_sent`**: inbound + outbound message hooks.
- **`session_start` / `session_end`**: session lifecycle boundaries.
- **`gateway_start` / `gateway_stop`**: gateway lifecycle events.
See [Plugins](/plugin#plugin-hooks) for the hook API and registration details.
## Streaming + partial replies
- Assistant deltas are streamed from pi-agent-core and emitted as `assistant` events.
- Block streaming can emit partial replies either on `text_end` or `message_end`.
- Reasoning streaming can be emitted as a separate stream or as block replies.
- See [Streaming](/concepts/streaming) for chunking and block reply behavior.
## Tool execution + messaging tools
- Tool start/update/end events are emitted on the `tool` stream.
- Tool results are sanitized for size and image payloads before logging/emitting.
- Messaging tool sends are tracked to suppress duplicate assistant confirmations.
## Reply shaping + suppression
- Final payloads are assembled from:
- assistant text (and optional reasoning)
- inline tool summaries (when verbose + allowed)
- assistant error text when the model errors
- `NO_REPLY` is treated as a silent token and filtered from outgoing payloads.
- Messaging tool duplicates are removed from the final payload list.
- If no renderable payloads remain and a tool errored, a fallback tool error reply is emitted
(unless a messaging tool already sent a user-visible reply).
## Compaction + retries
- Auto-compaction emits `compaction` stream events and can trigger a retry.
- On retry, in-memory buffers and tool summaries are reset to avoid duplicate output.
- See [Compaction](/concepts/compaction) for the compaction pipeline.
## Event streams (today)
- `lifecycle`: emitted by `subscribeEmbeddedPiSession` (and as a fallback by `agentCommand`)
- `assistant`: streamed deltas from pi-agent-core

View File

@ -86,6 +86,10 @@ These are the standard files Clawdbot expects inside the workspace:
- Optional tiny checklist for heartbeat runs.
- Keep it short to avoid token burn.
- `BOOT.md`
- Optional startup checklist executed on gateway restart when internal hooks are enabled.
- Keep it short; use the message tool for outbound sends.
- `BOOTSTRAP.md`
- One-time first-run ritual.
- Only created for a brand-new workspace.

View File

@ -21,7 +21,7 @@ Context is *not the same thing* as “memory”: memory can be stored on disk an
- `/status` → quick “how full is my window?” view + session settings.
- `/context list` → whats injected + rough sizes (per file + totals).
- `/context detail` → deeper breakdown: per-file, per-tool schema sizes, per-skill entry sizes, and system prompt size.
- `/cost on` → append per-reply usage line to normal replies.
- `/usage tokens` → append per-reply usage footer to normal replies.
- `/compact` → summarize older history into a compact entry to free window space.
See also: [Slash commands](/tools/slash-commands), [Token use & costs](/token-use), [Compaction](/concepts/compaction).
@ -149,4 +149,3 @@ Docs: [Session](/concepts/session), [Compaction](/concepts/compaction), [Session
- `System prompt (estimate)` = computed on the fly when no run report exists (or when running via a CLI backend that doesnt generate the report).
Either way, it reports sizes and top contributors; it does **not** dump the full system prompt or tool schemas.

View File

@ -9,6 +9,9 @@ read_when:
Clawdbot memory is **plain Markdown in the agent workspace**. The files are the
source of truth; the model only "remembers" what gets written to disk.
Memory search tools are provided by the active memory plugin (default:
`memory-core`). Disable memory plugins with `plugins.slots.memory = "none"`.
## Memory files (Markdown)
The default workspace layout uses two memory layers:
@ -76,18 +79,46 @@ semantic queries can find related notes even when wording differs.
Defaults:
- Enabled by default.
- Watches memory files for changes (debounced).
- Uses remote embeddings (OpenAI) unless configured for local.
- Uses remote embeddings by default. If `memorySearch.provider` is not set, Clawdbot auto-selects:
1. `local` if a `memorySearch.local.modelPath` is configured and the file exists.
2. `openai` if an OpenAI key can be resolved.
3. `gemini` if a Gemini key can be resolved.
4. Otherwise memory search stays disabled until configured.
- Local mode uses node-llama-cpp and may require `pnpm approve-builds`.
- Uses sqlite-vec (when available) to accelerate vector search inside SQLite.
Remote embeddings **require** an API key for the embedding provider. By default
this is OpenAI (`OPENAI_API_KEY` or `models.providers.openai.apiKey`). Codex
OAuth only covers chat/completions and does **not** satisfy embeddings for
memory search. When using a custom OpenAI-compatible endpoint, set
`memorySearch.remote.apiKey` (and optional `memorySearch.remote.headers`).
Remote embeddings **require** an API key for the embedding provider. Clawdbot
resolves keys from auth profiles, `models.providers.*.apiKey`, or environment
variables. Codex OAuth only covers chat/completions and does **not** satisfy
embeddings for memory search. For Gemini, use `GEMINI_API_KEY` or
`models.providers.google.apiKey`. When using a custom OpenAI-compatible endpoint,
set `memorySearch.remote.apiKey` (and optional `memorySearch.remote.headers`).
If you want to use a **custom OpenAI-compatible endpoint** (like Gemini, OpenRouter, or a proxy),
you can use the `remote` configuration:
### Gemini embeddings (native)
Set the provider to `gemini` to use the Gemini embeddings API directly:
```json5
agents: {
defaults: {
memorySearch: {
provider: "gemini",
model: "gemini-embedding-001",
remote: {
apiKey: "YOUR_GEMINI_API_KEY"
}
}
}
}
```
Notes:
- `remote.baseUrl` is optional (defaults to the Gemini API base URL).
- `remote.headers` lets you add extra headers if needed.
- Default model: `gemini-embedding-001`.
If you want to use a **custom OpenAI-compatible endpoint** (OpenRouter, vLLM, or a proxy),
you can use the `remote` configuration with the OpenAI provider:
```json5
agents: {
@ -96,8 +127,8 @@ agents: {
provider: "openai",
model: "text-embedding-3-small",
remote: {
baseUrl: "https://generativelanguage.googleapis.com/v1beta/openai/",
apiKey: "YOUR_GEMINI_API_KEY",
baseUrl: "https://api.example.com/v1/",
apiKey: "YOUR_OPENAI_COMPAT_API_KEY",
headers: { "X-Custom-Header": "value" }
}
}
@ -108,6 +139,24 @@ agents: {
If you don't want to set an API key, use `memorySearch.provider = "local"` or set
`memorySearch.fallback = "none"`.
Fallbacks:
- `memorySearch.fallback` can be `openai`, `gemini`, `local`, or `none`.
- The fallback provider is only used when the primary embedding provider fails.
Batch indexing (OpenAI + Gemini):
- Enabled by default for OpenAI and Gemini embeddings. Set `agents.defaults.memorySearch.remote.batch.enabled = false` to disable.
- Default behavior waits for batch completion; tune `remote.batch.wait`, `remote.batch.pollIntervalMs`, and `remote.batch.timeoutMinutes` if needed.
- Set `remote.batch.concurrency` to control how many batch jobs we submit in parallel (default: 2).
- Batch mode applies when `memorySearch.provider = "openai"` or `"gemini"` and uses the corresponding API key.
- Gemini batch jobs use the async embeddings batch endpoint and require Gemini Batch API availability.
Why OpenAI batch is fast + cheap:
- For large backfills, OpenAI is typically the fastest option we support because we can submit many embedding requests in a single batch job and let OpenAI process them asynchronously.
- OpenAI offers discounted pricing for Batch API workloads, so large indexing runs are usually cheaper than sending the same requests synchronously.
- See the OpenAI Batch API docs and pricing for details:
- https://platform.openai.com/docs/api-reference/batch
- https://platform.openai.com/pricing
Config example:
```json5
@ -117,6 +166,9 @@ agents: {
provider: "openai",
model: "text-embedding-3-small",
fallback: "openai",
remote: {
batch: { enabled: true, concurrency: 2 }
},
sync: { watch: true }
}
}
@ -141,8 +193,93 @@ Local mode:
### What gets indexed (and when)
- File type: Markdown only (`MEMORY.md`, `memory/**/*.md`).
- Index storage: per-agent SQLite at `~/.clawdbot/state/memory/<agentId>.sqlite` (configurable via `agents.defaults.memorySearch.store.path`, supports `{agentId}` token).
- Freshness: watcher on `MEMORY.md` + `memory/` marks the index dirty (debounce 1.5s). Sync runs on session start, on first search when dirty, and optionally on an interval. Reindex triggers when embedding model/provider or chunk sizes change.
- Index storage: per-agent SQLite at `~/.clawdbot/memory/<agentId>.sqlite` (configurable via `agents.defaults.memorySearch.store.path`, supports `{agentId}` token).
- Freshness: watcher on `MEMORY.md` + `memory/` marks the index dirty (debounce 1.5s). Sync runs on session start, on first search when dirty, and optionally on an interval.
- Reindex triggers: the index stores the embedding **provider/model + endpoint fingerprint + chunking params**. If any of those change, Clawdbot automatically resets and reindexes the entire store.
### Hybrid search (BM25 + vector)
When enabled, Clawdbot combines:
- **Vector similarity** (semantic match, wording can differ)
- **BM25 keyword relevance** (exact tokens like IDs, env vars, code symbols)
If full-text search is unavailable on your platform, Clawdbot falls back to vector-only search.
#### Why hybrid?
Vector search is great at “this means the same thing”:
- “Mac Studio gateway host” vs “the machine running the gateway”
- “debounce file updates” vs “avoid indexing on every write”
But it can be weak at exact, high-signal tokens:
- IDs (`a828e60`, `b3b9895a…`)
- code symbols (`memorySearch.query.hybrid`)
- error strings (“sqlite-vec unavailable”)
BM25 (full-text) is the opposite: strong at exact tokens, weaker at paraphrases.
Hybrid search is the pragmatic middle ground: **use both retrieval signals** so you get
good results for both “natural language” queries and “needle in a haystack” queries.
#### How we merge results (the current design)
Implementation sketch:
1) Retrieve a candidate pool from both sides:
- **Vector**: top `maxResults * candidateMultiplier` by cosine similarity.
- **BM25**: top `maxResults * candidateMultiplier` by FTS5 BM25 rank (lower is better).
2) Convert BM25 rank into a 0..1-ish score:
- `textScore = 1 / (1 + max(0, bm25Rank))`
3) Union candidates by chunk id and compute a weighted score:
- `finalScore = vectorWeight * vectorScore + textWeight * textScore`
Notes:
- `vectorWeight` + `textWeight` is normalized to 1.0 in config resolution, so weights behave as percentages.
- If embeddings are unavailable (or the provider returns a zero-vector), we still run BM25 and return keyword matches.
- If FTS5 cant be created, we keep vector-only search (no hard failure).
This isnt “IR-theory perfect”, but its simple, fast, and tends to improve recall/precision on real notes.
If we want to get fancier later, common next steps are Reciprocal Rank Fusion (RRF) or score normalization
(min/max or z-score) before mixing.
Config:
```json5
agents: {
defaults: {
memorySearch: {
query: {
hybrid: {
enabled: true,
vectorWeight: 0.7,
textWeight: 0.3,
candidateMultiplier: 4
}
}
}
}
}
```
### Embedding cache
Clawdbot can cache **chunk embeddings** in SQLite so reindexing and frequent updates (especially session transcripts) don't re-embed unchanged text.
Config:
```json5
agents: {
defaults: {
memorySearch: {
cache: {
enabled: true,
maxEntries: 50000
}
}
}
}
```
### Session memory search (experimental)

View File

@ -59,6 +59,11 @@ It does **not** rotate on every request. The pinned profile is reused until:
Manual selection via `/model …@<profileId>` sets a **user override** for that session
and is not autorotated until a new session starts.
Autopinned profiles (selected by the session router) are treated as a **preference**:
they are tried first, but Clawdbot may rotate to another profile on rate limits/timeouts.
Userpinned profiles stay locked to that profile; if it fails and model fallbacks
are configured, Clawdbot moves to the next model instead of switching profiles.
### Why OAuth can “look lost”
If you have both an OAuth profile and an API key profile for the same provider, roundrobin can switch between them across messages unless pinned. To force a single profile:

View File

@ -183,6 +183,22 @@ Kimi Code uses a dedicated endpoint and key (separate from Moonshot):
}
```
### Qwen OAuth (free tier)
Qwen provides OAuth access to Qwen Coder + Vision via a device-code flow.
Enable the bundled plugin, then log in:
```bash
clawdbot plugins enable qwen-portal-auth
clawdbot models auth login --provider qwen-portal --set-default
```
Model refs:
- `qwen-portal/coder-model`
- `qwen-portal/vision-model`
See [/providers/qwen](/providers/qwen) for setup details and notes.
### Synthetic
Synthetic provides Anthropic-compatible models behind the `synthetic` provider:

View File

@ -25,6 +25,7 @@ All session state is **owned by the gateway** (the “master” Clawdbot). UI cl
- Transcripts: `~/.clawdbot/agents/<agentId>/sessions/<SessionId>.jsonl` (Telegram topic sessions use `.../<SessionId>-topic-<threadId>.jsonl`).
- The store is a map `sessionKey -> { sessionId, updatedAt, ... }`. Deleting entries is safe; they are recreated on demand.
- Group entries may include `displayName`, `channel`, `subject`, `room`, and `space` to label sessions in UIs.
- Session entries include `origin` metadata (label + routing hints) so UIs can explain where a session came from.
- Clawdbot does **not** read legacy Pi/Tau session folders.
## Session pruning
@ -53,8 +54,12 @@ the workspace is writable. See [Memory](/concepts/memory) and
- Webhooks: `hook:<uuid>` (unless explicitly set by the hook)
- Node bridge runs: `node-<nodeId>`
## Lifecyle
- Idle expiry: `session.idleMinutes` (default 60). After the timeout a new `sessionId` is minted on the next message.
## Lifecycle
- Reset policy: sessions are reused until they expire, and expiry is evaluated on the next inbound message.
- Daily reset: defaults to **4:00 AM local time on the gateway host**. A session is stale once its last update is earlier than the most recent daily reset time.
- Idle reset (optional): `idleMinutes` adds a sliding idle window. When both daily and idle resets are configured, **whichever expires first** forces a new session.
- Legacy idle-only: if you set `session.idleMinutes` without any `session.reset`/`resetByType` config, Clawdbot stays in idle-only mode for backward compatibility.
- Per-type overrides (optional): `resetByType` lets you override the policy for `dm`, `group`, and `thread` sessions (thread = Slack/Discord threads, Telegram topics, Matrix threads when provided by the connector).
- Reset triggers: exact `/new` or `/reset` (plus any extras in `resetTriggers`) start a fresh session id and pass the remainder of the message through. If `/new` or `/reset` is sent alone, Clawdbot runs a short “hello” greeting turn to confirm the reset.
- Manual reset: delete specific keys from the store or remove the JSONL transcript; the next message recreates them.
- Isolated cron jobs always mint a fresh `sessionId` per run (no idle reuse).
@ -92,7 +97,18 @@ Send these as standalone messages so they register.
identityLinks: {
alice: ["telegram:123456789", "discord:987654321012345678"]
},
idleMinutes: 120,
reset: {
// Defaults: mode=daily, atHour=4 (gateway host local time).
// If you also set idleMinutes, whichever expires first wins.
mode: "daily",
atHour: 4,
idleMinutes: 120
},
resetByType: {
thread: { mode: "daily", atHour: 4 },
dm: { mode: "idle", idleMinutes: 240 },
group: { mode: "idle", idleMinutes: 120 }
},
resetTriggers: ["/new", "/reset"],
store: "~/.clawdbot/agents/{agentId}/sessions/sessions.json",
mainKey: "main",
@ -113,3 +129,18 @@ Send these as standalone messages so they register.
## Tips
- Keep the primary key dedicated to 1:1 traffic; let groups keep their own keys.
- When automating cleanup, delete individual keys instead of the whole store to preserve context elsewhere.
## Session origin metadata
Each session entry records where it came from (best-effort) in `origin`:
- `label`: human label (resolved from conversation label + group subject/channel)
- `provider`: normalized channel id (including extensions)
- `from`/`to`: raw routing ids from the inbound envelope
- `accountId`: provider account id (when multi-account)
- `threadId`: thread/topic id when the channel supports it
The origin fields are populated for direct messages, channels, and groups. If a
connector only updates delivery routing (for example, to keep a DM main session
fresh), it should still provide inbound context so the session keeps its
explainer metadata. Extensions can do this by sending `ConversationLabel`,
`GroupSubject`, `GroupChannel`, `GroupSpace`, and `SenderName` in the inbound
context and calling `recordSessionMetaFromInbound` (or passing the same context
to `updateLastRoute`).

View File

@ -18,6 +18,7 @@ The prompt is intentionally compact and uses fixed sections:
- **Skills** (when available): tells the model how to load skill instructions on demand.
- **Clawdbot Self-Update**: how to run `config.apply` and `update.run`.
- **Workspace**: working directory (`agents.defaults.workspace`).
- **Documentation**: local path to Clawdbot docs (repo or npm package) and when to read them.
- **Workspace Files (injected)**: indicates bootstrap files are included below.
- **Sandbox** (when enabled): indicates sandboxed runtime, sandbox paths, and whether elevated exec is available.
- **Current Date & Time**: user-local time, timezone, and time format.
@ -58,6 +59,9 @@ Large files are truncated with a marker. The max per-file size is controlled by
`agents.defaults.bootstrapMaxChars` (default: 20000). Missing files inject a
short missing-file marker.
Internal hooks can intercept this step via `agent:bootstrap` to mutate or replace
the injected bootstrap files (for example swapping `SOUL.md` for an alternate persona).
To inspect how much each injected file contributes (raw vs injected, truncation, plus tool schema overhead), use `/context list` or `/context detail`. See [Context](/concepts/context).
## Time handling
@ -95,3 +99,12 @@ Skills section is omitted.
```
This keeps the base prompt small while still enabling targeted skill usage.
## Documentation
When available, the system prompt includes a **Documentation** section that points to the
local Clawdbot docs directory (either `docs/` in the repo workspace or the bundled npm
package docs) and also notes the public mirror, source repo, community Discord, and
ClawdHub (https://clawdhub.com) for skills discovery. The prompt instructs the model to consult local docs first
for Clawdbot behavior, commands, configuration, or architecture, and to run
`clawdbot status` itself when possible (asking the user only when it lacks access).

View File

@ -12,7 +12,7 @@ read_when:
## Where it shows up
- `/status` in chats: emojirich status card with session tokens + estimated cost (API key only). Provider usage shows for the **current model provider** when available.
- `/cost on|off` in chats: toggles perresponse usage lines (OAuth shows tokens only).
- `/usage off|tokens|full` in chats: per-response usage footer (OAuth shows tokens only).
- CLI: `clawdbot status --usage` prints a full per-provider breakdown.
- CLI: `clawdbot channels list` prints the same usage snapshot alongside provider config (use `--no-usage` to skip).
- macOS menu bar: “Usage” section under Context (only if available).

View File

@ -77,6 +77,7 @@ What this does:
- Seeds the workspace files if missing:
`AGENTS.md`, `SOUL.md`, `TOOLS.md`, `IDENTITY.md`, `USER.md`, `HEARTBEAT.md`.
- Default identity: **C3PO** (protocol droid).
- Skips channel providers in dev mode (`CLAWDBOT_SKIP_CHANNELS=1`).
Reset flow (fresh start):

View File

@ -657,6 +657,10 @@
"source": "/templates/AGENTS",
"destination": "/reference/templates/AGENTS"
},
{
"source": "/templates/BOOT",
"destination": "/reference/templates/BOOT"
},
{
"source": "/templates/BOOTSTRAP",
"destination": "/reference/templates/BOOTSTRAP"
@ -822,8 +826,10 @@
"cli/models",
"cli/logs",
"cli/nodes",
"cli/approvals",
"cli/gateway",
"cli/daemon",
"cli/service",
"cli/tui",
"cli/voicecall",
"cli/wake",
@ -956,6 +962,8 @@
{
"group": "Automation & Hooks",
"pages": [
"hooks",
"hooks/soul-evil",
"automation/auth-monitoring",
"automation/webhook",
"automation/gmail-pubsub",
@ -1049,6 +1057,7 @@
"reference/RELEASING",
"reference/AGENTS.default",
"reference/templates/AGENTS",
"reference/templates/BOOT",
"reference/templates/BOOTSTRAP",
"reference/templates/HEARTBEAT",
"reference/templates/IDENTITY",

View File

@ -46,8 +46,8 @@ When TLS is enabled, discovery TXT records include `bridgeTls=1` plus
## Frames
Client → Gateway:
- `req` / `res`: scoped gateway RPC (chat, sessions, config, health, voicewake)
- `event`: node signals (voice transcript, agent request, chat subscribe)
- `req` / `res`: scoped gateway RPC (chat, sessions, config, health, voicewake, skills.bins)
- `event`: node signals (voice transcript, agent request, chat subscribe, exec lifecycle)
Gateway → Client:
- `invoke` / `invoke-res`: node commands (`canvas.*`, `camera.*`, `screen.record`,
@ -57,6 +57,18 @@ Gateway → Client:
Exact allowlist is enforced in `src/gateway/server-bridge.ts`.
## Exec lifecycle events
Nodes can emit `exec.started`, `exec.finished`, or `exec.denied` events to surface
system.run activity. These are mapped to system events in the gateway.
Payload fields (all optional unless noted):
- `sessionKey` (required): agent session to receive the system event.
- `runId`: unique exec id for grouping.
- `command`: raw or formatted command string.
- `exitCode`, `timedOut`, `success`, `output`: completion details (finished only).
- `reason`: denial reason (denied only).
## Tailnet usage
- Bind the bridge to a tailnet IP: `bridge.bind: "tailnet"` in

View File

@ -146,7 +146,11 @@ Save to `~/.clawdbot/clawdbot.json` and you can DM the bot from that number.
// Session behavior
session: {
scope: "per-sender",
idleMinutes: 60,
reset: {
mode: "daily",
atHour: 4,
idleMinutes: 60
},
heartbeatIdleMinutes: 120,
resetTriggers: ["/new", "/reset"],
store: "~/.clawdbot/agents/default/sessions/sessions.json",
@ -257,10 +261,9 @@ Save to `~/.clawdbot/clawdbot.json` and you can DM the bot from that number.
ackMaxChars: 300
},
memorySearch: {
provider: "openai",
model: "text-embedding-004",
provider: "gemini",
model: "gemini-embedding-001",
remote: {
baseUrl: "https://generativelanguage.googleapis.com/v1beta/openai/",
apiKey: "${GEMINI_API_KEY}"
}
},

View File

@ -678,10 +678,11 @@ Notes:
- `"open"`: groups bypass allowlists; mention-gating still applies.
- `"disabled"`: block all group/room messages.
- `"allowlist"`: only allow groups/rooms that match the configured allowlist.
- `channels.defaults.groupPolicy` sets the default when a providers `groupPolicy` is unset.
- WhatsApp/Telegram/Signal/iMessage/Microsoft Teams use `groupAllowFrom` (fallback: explicit `allowFrom`).
- Discord/Slack use channel allowlists (`channels.discord.guilds.*.channels`, `channels.slack.channels`).
- Group DMs (Discord/Slack) are still controlled by `dm.groupEnabled` + `dm.groupChannels`.
- Default is `groupPolicy: "allowlist"`; if no allowlist is configured, group messages are blocked.
- Default is `groupPolicy: "allowlist"` (unless overridden by `channels.defaults.groupPolicy`); if no allowlist is configured, group messages are blocked.
### Multi-agent routing (`agents.list` + `bindings`)
@ -2415,7 +2416,7 @@ Notes:
### `session`
Controls session scoping, idle expiry, reset triggers, and where the session store is written.
Controls session scoping, reset policy, reset triggers, and where the session store is written.
```json5
{
@ -2425,7 +2426,16 @@ Controls session scoping, idle expiry, reset triggers, and where the session sto
identityLinks: {
alice: ["telegram:123456789", "discord:987654321012345678"]
},
idleMinutes: 60,
reset: {
mode: "daily",
atHour: 4,
idleMinutes: 60
},
resetByType: {
thread: { mode: "daily", atHour: 4 },
dm: { mode: "idle", idleMinutes: 240 },
group: { mode: "idle", idleMinutes: 120 }
},
resetTriggers: ["/new", "/reset"],
// Default is already per-agent under ~/.clawdbot/agents/<agentId>/sessions/sessions.json
// You can override with {agentId} templating:
@ -2436,12 +2446,12 @@ Controls session scoping, idle expiry, reset triggers, and where the session sto
// Max ping-pong reply turns between requester/target (05).
maxPingPongTurns: 5
},
sendPolicy: {
rules: [
sendPolicy: {
rules: [
{ action: "deny", match: { channel: "discord", chatType: "group" } }
],
default: "allow"
}
],
default: "allow"
}
}
}
```
@ -2455,6 +2465,13 @@ Fields:
- `per-channel-peer`: isolate DMs per channel + sender (recommended for multi-user inboxes).
- `identityLinks`: map canonical ids to provider-prefixed peers so the same person shares a DM session across channels when using `per-peer` or `per-channel-peer`.
- Example: `alice: ["telegram:123456789", "discord:987654321012345678"]`.
- `reset`: primary reset policy. Defaults to daily resets at 4:00 AM local time on the gateway host.
- `mode`: `daily` or `idle` (default: `daily` when `reset` is present).
- `atHour`: local hour (0-23) for the daily reset boundary.
- `idleMinutes`: sliding idle window in minutes. When daily + idle are both configured, whichever expires first wins.
- `resetByType`: per-session overrides for `dm`, `group`, and `thread`.
- If you only set legacy `session.idleMinutes` without any `reset`/`resetByType`, Clawdbot stays in idle-only mode for backward compatibility.
- `heartbeatIdleMinutes`: optional idle override for heartbeat checks (daily reset still applies when enabled).
- `agentToAgent.maxPingPongTurns`: max reply-back turns between requester/target (05, default 5).
- `sendPolicy.default`: `allow` or `deny` fallback when no rule matches.
- `sendPolicy.rules[]`: match by `channel`, `chatType` (`direct|group|room`), or `keyPrefix` (e.g. `cron:`). First deny wins; otherwise allow.

View File

@ -65,8 +65,8 @@ stronger isolation between agents, run them under separate OS users or separate
If a macOS node is paired, the Gateway can invoke `system.run` on that node. This is **remote code execution** on the Mac:
- Requires node pairing (approval + token).
- Controlled on the Mac via **Settings → "Node Run Commands"**: "Always Ask" (default), "Always Allow", or "Never".
- If you dont want remote execution, set the policy to "Never" and remove node pairing for that Mac.
- Controlled on the Mac via **Settings → Exec approvals** (security + ask + allowlist).
- If you dont want remote execution, set security to **deny** and remove node pairing for that Mac.
## Dynamic skills (watcher / remote nodes)

View File

@ -239,11 +239,15 @@ Known issue: When you send an image with ONLY a mention (no other text), WhatsAp
ls -la ~/.clawdbot/agents/<agentId>/sessions/
```
**Check 2:** Is `idleMinutes` too short?
**Check 2:** Is the reset window too short?
```json
{
"session": {
"idleMinutes": 10080 // 7 days
"reset": {
"mode": "daily",
"atHour": 4,
"idleMinutes": 10080 // 7 days
}
}
}
```

View File

@ -14,6 +14,8 @@ Hooks are small scripts that run when something happens. There are two kinds:
- **Hooks** (this page): run inside the Gateway when agent events fire, like `/new`, `/reset`, `/stop`, or lifecycle events.
- **Webhooks**: external HTTP webhooks that let other systems trigger work in Clawdbot. See [Webhook Hooks](/automation/webhook) or use `clawdbot webhooks` for Gmail helper commands.
Hooks can also be bundled inside plugins; see [Plugins](/plugin#plugin-hooks).
Common uses:
- Save a memory snapshot when you reset a session
@ -35,10 +37,12 @@ The hooks system allows you to:
### Bundled Hooks
Clawdbot ships with two bundled hooks that are automatically discovered:
Clawdbot ships with four bundled hooks that are automatically discovered:
- **💾 session-memory**: Saves session context to your agent workspace (default `~/clawd/memory/`) when you issue `/new`
- **📝 command-logger**: Logs all command events to `~/.clawdbot/logs/commands.log`
- **🚀 boot-md**: Runs `BOOT.md` when the gateway starts (requires internal hooks enabled)
- **😈 soul-evil**: Swaps injected `SOUL.md` content with `SOUL_EVIL.md` during a purge window or by random chance
List available hooks:
@ -192,7 +196,7 @@ Each event includes:
```typescript
{
type: 'command' | 'session' | 'agent',
type: 'command' | 'session' | 'agent' | 'gateway',
action: string, // e.g., 'new', 'reset', 'stop'
sessionKey: string, // Session identifier
timestamp: Date, // When the event occurred
@ -203,6 +207,8 @@ Each event includes:
sessionFile?: string,
commandSource?: string, // e.g., 'whatsapp', 'telegram'
senderId?: string,
workspaceDir?: string,
bootstrapFiles?: WorkspaceBootstrapFile[],
cfg?: ClawdbotConfig
}
}
@ -219,6 +225,16 @@ Triggered when agent commands are issued:
- **`command:reset`**: When `/reset` command is issued
- **`command:stop`**: When `/stop` command is issued
### Agent Events
- **`agent:bootstrap`**: Before workspace bootstrap files are injected (hooks may mutate `context.bootstrapFiles`)
### Gateway Events
Triggered when the gateway starts:
- **`gateway:startup`**: After channels start and hooks are loaded
### Future Events
Planned event types:
@ -497,6 +513,62 @@ grep '"action":"new"' ~/.clawdbot/logs/commands.log | jq .
clawdbot hooks enable command-logger
```
### soul-evil
Swaps injected `SOUL.md` content with `SOUL_EVIL.md` during a purge window or by random chance.
**Events**: `agent:bootstrap`
**Docs**: [SOUL Evil Hook](/hooks/soul-evil)
**Output**: No files written; swaps happen in-memory only.
**Enable**:
```bash
clawdbot hooks enable soul-evil
```
**Config**:
```json
{
"hooks": {
"internal": {
"enabled": true,
"entries": {
"soul-evil": {
"enabled": true,
"file": "SOUL_EVIL.md",
"chance": 0.1,
"purge": { "at": "21:00", "duration": "15m" }
}
}
}
}
}
```
### boot-md
Runs `BOOT.md` when the gateway starts (after channels start).
Internal hooks must be enabled for this to run.
**Events**: `gateway:startup`
**Requirements**: `workspace.dir` must be configured
**What it does**:
1. Reads `BOOT.md` from your workspace
2. Runs the instructions via the agent runner
3. Sends any requested outbound messages via the message tool
**Enable**:
```bash
clawdbot hooks enable boot-md
```
## Best Practices
### Keep Handlers Fast
@ -569,6 +641,7 @@ The gateway logs hook loading at startup:
```
Registered hook: session-memory -> command:new
Registered hook: command-logger -> command
Registered hook: boot-md -> gateway:startup
```
### Check Discovery

68
docs/hooks/soul-evil.md Normal file
View File

@ -0,0 +1,68 @@
---
summary: "SOUL Evil hook (swap SOUL.md with SOUL_EVIL.md)"
read_when:
- You want to enable or tune the SOUL Evil hook
- You want a purge window or random-chance persona swap
---
# SOUL Evil Hook
The SOUL Evil hook swaps the **injected** `SOUL.md` content with `SOUL_EVIL.md` during
a purge window or by random chance. It does **not** modify files on disk.
## How It Works
When `agent:bootstrap` runs, the hook can replace the `SOUL.md` content in memory
before the system prompt is assembled. If `SOUL_EVIL.md` is missing or empty,
Clawdbot logs a warning and keeps the normal `SOUL.md`.
Sub-agent runs do **not** include `SOUL.md` in their bootstrap files, so this hook
has no effect on sub-agents.
## Enable
```bash
clawdbot hooks enable soul-evil
```
Then set the config:
```json
{
"hooks": {
"internal": {
"enabled": true,
"entries": {
"soul-evil": {
"enabled": true,
"file": "SOUL_EVIL.md",
"chance": 0.1,
"purge": { "at": "21:00", "duration": "15m" }
}
}
}
}
}
```
Create `SOUL_EVIL.md` in the agent workspace root (next to `SOUL.md`).
## Options
- `file` (string): alternate SOUL filename (default: `SOUL_EVIL.md`)
- `chance` (number 01): random chance per run to use `SOUL_EVIL.md`
- `purge.at` (HH:mm): daily purge start (24-hour clock)
- `purge.duration` (duration): window length (e.g. `30s`, `10m`, `1h`)
**Precedence:** purge window wins over chance.
**Timezone:** uses `agents.defaults.userTimezone` when set; otherwise host timezone.
## Notes
- No files are written or modified on disk.
- If `SOUL.md` is not in the bootstrap list, the hook does nothing.
## See Also
- [Hooks](/hooks)

View File

@ -12,6 +12,7 @@ Clawdbot ships two installer scripts (served from `clawd.bot`):
- `https://clawd.bot/install.sh` — “recommended” installer (global npm install by default; can also install from a GitHub checkout)
- `https://clawd.bot/install-cli.sh` — non-root-friendly CLI installer (installs into a prefix with its own Node)
- `https://clawd.bot/install.ps1` — Windows PowerShell installer (npm by default; optional git install)
To see the current flags/behavior, run:
@ -19,6 +20,12 @@ To see the current flags/behavior, run:
curl -fsSL https://clawd.bot/install.sh | bash -s -- --help
```
Windows (PowerShell) help:
```powershell
& ([scriptblock]::Create((iwr -useb https://clawd.bot/install.ps1))) -?
```
If the installer completes but `clawdbot` is not found in a new terminal, its usually a Node/npm PATH issue. See: [Install](/install#nodejs--npm-path-sanity).
## install.sh (recommended)
@ -73,3 +80,37 @@ Help:
```bash
curl -fsSL https://clawd.bot/install-cli.sh | bash -s -- --help
```
## install.ps1 (Windows PowerShell)
What it does (high level):
- Ensure Node.js **22+** (winget/Chocolatey/Scoop or manual).
- Choose install method:
- `npm` (default): `npm install -g clawdbot@latest`
- `git`: clone/build a source checkout and install a wrapper script
- Runs `clawdbot doctor --non-interactive` on upgrades and git installs (best effort).
Examples:
```powershell
iwr -useb https://clawd.bot/install.ps1 | iex
```
```powershell
iwr -useb https://clawd.bot/install.ps1 | iex -InstallMethod git
```
```powershell
iwr -useb https://clawd.bot/install.ps1 | iex -InstallMethod git -GitDir "C:\\clawdbot"
```
Environment variables:
- `CLAWDBOT_INSTALL_METHOD=git|npm`
- `CLAWDBOT_GIT_DIR=...`
Git requirement:
If you choose `-InstallMethod git` and Git is missing, the installer will print the
Git for Windows link (`https://git-scm.com/download/win`) and exit.

View File

@ -147,9 +147,10 @@ Notes:
- The permission prompt must be accepted on the Android device before the capability is advertised.
- Wi-Fi-only devices without telephony will not advertise `sms.send`.
## System commands (mac node)
## System commands (node host / mac node)
The macOS node exposes `system.run` and `system.notify`.
The macOS node exposes `system.run`, `system.notify`, and `system.execApprovals.get/set`.
The headless node host exposes `system.run`, `system.which`, and `system.execApprovals.get/set`.
Examples:
@ -163,12 +164,58 @@ Notes:
- `system.notify` respects notification permission state on the macOS app.
- `system.run` supports `--cwd`, `--env KEY=VAL`, `--command-timeout`, and `--needs-screen-recording`.
- `system.notify` supports `--priority <passive|active|timeSensitive>` and `--delivery <system|overlay|auto>`.
- `system.run` is gated by the macOS app policy (Settings → "Node Run Commands"): "Always Ask" prompts per command, "Always Allow" runs without prompts, and "Never" disables the tool. Denied prompts return `SYSTEM_RUN_DENIED`; disabled returns `SYSTEM_RUN_DISABLED`.
- On macOS node mode, `system.run` is gated by exec approvals in the macOS app (Settings → Exec approvals).
Ask/allowlist/full behave the same as the headless node host; denied prompts return `SYSTEM_RUN_DENIED`.
- On headless node host, `system.run` is gated by exec approvals (`~/.clawdbot/exec-approvals.json`).
## Exec node binding
When multiple nodes are available, you can bind exec to a specific node.
This sets the default node for `exec host=node` (and can be overridden per agent).
Global default:
```bash
clawdbot config set tools.exec.node "node-id-or-name"
```
Per-agent override:
```bash
clawdbot config get agents.list
clawdbot config set agents.list[0].tools.exec.node "node-id-or-name"
```
Unset to allow any node:
```bash
clawdbot config unset tools.exec.node
clawdbot config unset agents.list[0].tools.exec.node
```
## Permissions map
Nodes may include a `permissions` map in `node.list` / `node.describe`, keyed by permission name (e.g. `screenRecording`, `accessibility`) with boolean values (`true` = granted).
## Headless node host (cross-platform)
Clawdbot can run a **headless node host** (no UI) that connects to the Gateway
bridge and exposes `system.run` / `system.which`. This is useful on Linux/Windows
or for running a minimal node alongside a server.
Start it:
```bash
clawdbot node start --host <gateway-host> --port 18790
```
Notes:
- Pairing is still required (the Gateway will show a node approval prompt).
- The node host stores its node id + pairing token in `~/.clawdbot/node.json`.
- Exec approvals are enforced locally via `~/.clawdbot/exec-approvals.json`
(see [Exec approvals](/tools/exec-approvals)).
- Add `--tls` / `--tls-fingerprint` when the bridge requires TLS.
## Mac node mode
- The macOS menubar app connects to the Gateway bridge as a node (so `clawdbot nodes …` works against this Mac).

View File

@ -104,6 +104,29 @@ Rules:
- If `<capability>.enabled: true` but no models are configured, Clawdbot tries the
**active reply model** when its provider supports the capability.
### Auto-enable audio (when keys exist)
If `tools.media.audio.enabled` is **not** set to `false` and you have any supported
audio provider keys configured, Clawdbot will **auto-enable audio transcription**
even when you havent listed models explicitly.
Providers checked (in order):
1) OpenAI
2) Groq
3) Deepgram
To disable this behavior, set:
```json5
{
tools: {
media: {
audio: {
enabled: false
}
}
}
}
```
## Capabilities (optional)
If you set `capabilities`, the entry only runs for those media types. For shared
lists, Clawdbot can infer defaults:

76
docs/perplexity.md Normal file
View File

@ -0,0 +1,76 @@
---
summary: "Perplexity Sonar setup for web_search"
read_when:
- You want to use Perplexity Sonar for web search
- You need PERPLEXITY_API_KEY or OpenRouter setup
---
# Perplexity Sonar
Clawdbot can use Perplexity Sonar for the `web_search` tool. You can connect
through Perplexitys direct API or via OpenRouter.
## API options
### Perplexity (direct)
- Base URL: https://api.perplexity.ai
- Environment variable: `PERPLEXITY_API_KEY`
### OpenRouter (alternative)
- Base URL: https://openrouter.ai/api/v1
- Environment variable: `OPENROUTER_API_KEY`
- Supports prepaid/crypto credits.
## Config example
```json5
{
tools: {
web: {
search: {
provider: "perplexity",
perplexity: {
apiKey: "pplx-...",
baseUrl: "https://api.perplexity.ai",
model: "perplexity/sonar-pro"
}
}
}
}
}
```
## Switching from Brave
```json5
{
tools: {
web: {
search: {
provider: "perplexity",
perplexity: {
apiKey: "pplx-...",
baseUrl: "https://api.perplexity.ai"
}
}
}
}
}
```
If both `PERPLEXITY_API_KEY` and `OPENROUTER_API_KEY` are set, set
`tools.web.search.perplexity.baseUrl` (or `tools.web.search.perplexity.apiKey`)
to disambiguate.
If `PERPLEXITY_API_KEY` is used from the environment and no base URL is set,
Clawdbot defaults to the direct Perplexity endpoint. Set `baseUrl` to override.
## Models
- `perplexity/sonar` — fast Q&A with web search
- `perplexity/sonar-pro` (default) — multi-step reasoning + web search
- `perplexity/sonar-reasoning-pro` — deep research
See [Web tools](/tools/web) for the full web_search configuration.

View File

@ -9,8 +9,9 @@ read_when:
# Gateway on macOS (external launchd)
Clawdbot.app no longer bundles Node/Bun or the Gateway runtime. The macOS app
expects an **external** `clawdbot` CLI install and manages a peruser launchd
service to keep the Gateway running.
expects an **external** `clawdbot` CLI install, does not spawn the Gateway as a
child process, and manages a peruser launchd service to keep the Gateway
running (or attaches to an existing local Gateway if one is already running).
## Install the CLI (required for local mode)
@ -38,6 +39,8 @@ Manager:
Behavior:
- “Clawdbot Active” enables/disables the LaunchAgent.
- App quit does **not** stop the gateway (launchd keeps it alive).
- If a Gateway is already running on the configured port, the app attaches to
it instead of starting a new one.
Logging:
- launchd stdout/err: `/tmp/clawdbot/clawdbot-gateway.log`

View File

@ -5,9 +5,11 @@ read_when:
---
# Gateway lifecycle on macOS
The macOS app **manages the Gateway via launchd** by default. The launchd job
uses the external `clawdbot` CLI (no embedded runtime). This gives you reliable
autostart at login and restart on crashes.
The macOS app **manages the Gateway via launchd** by default and does not spawn
the Gateway as a child process. It first tries to attach to an alreadyrunning
Gateway on the configured port; if none is reachable, it enables the launchd
service via the external `clawdbot` CLI (no embedded runtime). This gives you
reliable autostart at login and restart on crashes.
Childprocess mode (Gateway spawned directly by the app) is **not in use** today.
If you need tighter coupling to the UI, run the Gateway manually in a terminal.

View File

@ -14,15 +14,7 @@ Before building the app, ensure you have the following installed:
1. **Xcode 26.2+**: Required for Swift development.
2. **Node.js 22+ & pnpm**: Required for the gateway, CLI, and packaging scripts.
## 1. Initialize Submodules
Clawdbot depends on several submodules (like `Peekaboo`). You must initialize these recursively:
```bash
git submodule update --init --recursive
```
## 2. Install Dependencies
## 1. Install Dependencies
Install the project-wide dependencies:
@ -30,7 +22,7 @@ Install the project-wide dependencies:
pnpm install
```
## 3. Build and Package the App
## 2. Build and Package the App
To build the macOS app and package it into `dist/Clawdbot.app`, run:
@ -40,9 +32,12 @@ To build the macOS app and package it into `dist/Clawdbot.app`, run:
If you don't have an Apple Developer ID certificate, the script will automatically use **ad-hoc signing** (`-`).
For dev run modes, signing flags, and Team ID troubleshooting, see the macOS app README:
https://github.com/clawdbot/clawdbot/blob/main/apps/macos/README.md
> **Note**: Ad-hoc signed apps may trigger security prompts. If the app crashes immediately with "Abort trap 6", see the [Troubleshooting](#troubleshooting) section.
## 4. Install the CLI
## 3. Install the CLI
The macOS app expects a global `clawdbot` CLI install to manage background tasks.

View File

@ -2,7 +2,7 @@
summary: "PeekabooBridge integration for macOS UI automation"
read_when:
- Hosting PeekabooBridge in Clawdbot.app
- Integrating Peekaboo as a submodule
- Integrating Peekaboo via Swift Package Manager
- Changing PeekabooBridge protocol/paths
---
# Peekaboo Bridge (macOS UI automation)

View File

@ -14,6 +14,7 @@ This app is usually built from [`scripts/package-mac-app.sh`](https://github.com
- inject build metadata into Info.plist: `ClawdbotBuildTimestamp` (UTC) and `ClawdbotGitCommit` (short hash) so the About pane can show build, git, and debug/release channel.
- **Packaging requires Node 22+**: the script runs TS builds and the Control UI build.
- reads `SIGN_IDENTITY` from the environment. Add `export SIGN_IDENTITY="Apple Development: Your Name (TEAMID)"` (or your Developer ID Application cert) to your shell rc to always sign with your cert. Ad-hoc signing requires explicit opt-in via `ALLOW_ADHOC_SIGNING=1` or `SIGN_IDENTITY="-"` (not recommended for permission testing).
- runs a Team ID audit after signing and fails if any Mach-O inside the app bundle is signed by a different Team ID. Set `SKIP_TEAM_ID_CHECK=1` to bypass.
## Usage
@ -23,6 +24,7 @@ scripts/package-mac-app.sh # auto-selects identity; errors if none
SIGN_IDENTITY="Developer ID Application: Your Name" scripts/package-mac-app.sh # real cert
ALLOW_ADHOC_SIGNING=1 scripts/package-mac-app.sh # ad-hoc (permissions will not stick)
SIGN_IDENTITY="-" scripts/package-mac-app.sh # explicit ad-hoc (same caveat)
DISABLE_LIBRARY_VALIDATION=1 scripts/package-mac-app.sh # dev-only Sparkle Team ID mismatch workaround
```
### Ad-hoc Signing Note

View File

@ -7,6 +7,8 @@ read_when:
**Current model:** there is **no local control socket** and no `clawdbot-mac` CLI. All agent actions go through the Gateway WebSocket and `node.invoke`. UI automation still uses PeekabooBridge.
**Planned model:** add a local Unix socket between the **node service** and the **macOS app**. The app owns `system.run` (UI/TCC context); the node service forwards exec requests over IPC.
## Goals
- Single GUI app instance that owns all TCC-facing work (notifications, screen recording, mic, speech, AppleScript).
- A small surface for automation: Gateway + node commands, plus PeekabooBridge for UI automation.
@ -17,6 +19,19 @@ read_when:
- The app runs the Gateway (local mode) and connects to it as a node.
- Agent actions are performed via `node.invoke` (e.g. `system.run`, `system.notify`, `canvas.*`).
### Node service + app IPC (planned)
- A headless node service connects to the Gateway bridge.
- `system.run` requests are forwarded to the macOS app over a local Unix socket.
- The app performs the exec in UI context, prompts if needed, and returns output.
Diagram (SCI):
```
Agent -> Gateway -> Bridge -> Node Service (TS)
| IPC (UDS + token + HMAC + TTL)
v
Mac App (UI + TCC + system.run)
```
### PeekabooBridge (UI automation)
- UI automation uses a separate UNIX socket named `bridge.sock` and the PeekabooBridge JSON protocol.
- Host preference order (client-side): Peekaboo.app → Claude.app → Clawdbot.app → local execution.
@ -25,6 +40,7 @@ read_when:
### Mach/XPC
- Not required for automation; `node.invoke` + PeekabooBridge cover current needs.
- Planned IPC keeps Unix sockets (no XPC helper).
## Operational flows
- Restart/rebuild: `SIGN_IDENTITY="Apple Development: <Developer Name> (<TEAMID>)" scripts/restart-mac.sh`
@ -38,3 +54,4 @@ read_when:
- PeekabooBridge: `PEEKABOO_ALLOW_UNSIGNED_SOCKET_CLIENTS=1` (DEBUG-only) may allow same-UID callers for local development.
- All communication remains local-only; no network sockets are exposed.
- TCC prompts originate only from the GUI app bundle; keep the signed bundle ID stable across rebuilds.
- Planned IPC hardening: socket mode `0600`, token, peer-UID checks, HMAC challenge/response, short TTL.

View File

@ -7,8 +7,8 @@ read_when:
# Clawdbot macOS Companion (menu bar + gateway broker)
The macOS app is the **menubar companion** for Clawdbot. It owns permissions,
manages the Gateway locally, and exposes macOS capabilities to the agent as a
node.
manages/attaches to the Gateway locally (launchd or manual), and exposes macOS
capabilities to the agent as a node.
## What it does
@ -20,13 +20,17 @@ node.
- Optionally hosts **PeekabooBridge** for UI automation.
- Installs the global CLI (`clawdbot`) via npm/pnpm on request (bun not recommended for the Gateway runtime).
Planned:
- Run a headless **node service** locally (launchd).
- Keep `system.run` in the app (UI/TCC context), with the node service forwarding via IPC.
## Local vs remote mode
- **Local** (default): the app ensures a local Gateway is running via launchd.
- **Local** (default): the app attaches to a running local Gateway if present;
otherwise it enables the launchd service via `clawdbot daemon`.
- **Remote**: the app connects to a Gateway over SSH/Tailscale and never starts
a local process.
- **Attachonly** (debug): the app connects to an alreadyrunning local Gateway
and never spawns its own.
The app does not spawn the Gateway as a child process.
## Launchd control
@ -54,29 +58,44 @@ The macOS app presents itself as a node. Common commands:
The node reports a `permissions` map so agents can decide whats allowed.
## Node run policy + allowlist
`system.run` is controlled by the macOS app **Node Run Commands** policy:
- `Always Ask`: prompt per command (default).
- `Always Allow`: run without prompts.
- `Never`: disable `system.run` (tool not advertised).
The policy + allowlist live on the Mac in:
Planned split:
- Node service advertises the node surface to the Gateway.
- macOS app performs `system.run` in UI context over IPC.
Diagram (SCI):
```
~/.clawdbot/macos-node.json
Gateway -> Bridge -> Node Service (TS)
| IPC (UDS + token + HMAC + TTL)
v
Mac App (UI + TCC + system.run)
```
Schema:
## Exec approvals (system.run)
`system.run` is controlled by **Exec approvals** in the macOS app (Settings → Exec approvals).
Security + ask + allowlist are stored locally on the Mac in:
```
~/.clawdbot/exec-approvals.json
```
Example:
```json
{
"systemRun": {
"policy": "ask",
"allowlist": [
"[\"/bin/echo\",\"hello\"]"
]
"version": 1,
"defaults": {
"security": "deny",
"ask": "on-miss"
},
"agents": {
"main": {
"security": "allowlist",
"ask": "on-miss",
"allowlist": [
{ "pattern": "/opt/homebrew/bin/rg" }
]
}
}
}
```

View File

@ -36,6 +36,8 @@ See [Voice Call](/plugins/voice-call) for a concrete example plugin.
## Available plugins (official)
- Microsoft Teams is plugin-only as of 2026.1.15; install `@clawdbot/msteams` if you use Teams.
- Memory (Core) — bundled memory search plugin (enabled by default via `plugins.slots.memory`)
- Memory (LanceDB) — bundled long-term memory plugin (auto-recall/capture; set `plugins.slots.memory = "memory-lancedb"`)
- [Voice Call](/plugins/voice-call) — `@clawdbot/voice-call`
- [Zalo Personal](/plugins/zalouser) — `@clawdbot/zalouser`
- [Matrix](/channels/matrix) — `@clawdbot/matrix`
@ -43,7 +45,8 @@ See [Voice Call](/plugins/voice-call) for a concrete example plugin.
- [Microsoft Teams](/channels/msteams) — `@clawdbot/msteams`
- Google Antigravity OAuth (provider auth) — bundled as `google-antigravity-auth` (disabled by default)
- Gemini CLI OAuth (provider auth) — bundled as `google-gemini-cli-auth` (disabled by default)
- Copilot Proxy (provider auth) — bundled as `copilot-proxy` (disabled by default)
- Qwen OAuth (provider auth) — bundled as `qwen-portal-auth` (disabled by default)
- Copilot Proxy (provider auth) — local VS Code Copilot Proxy bridge; distinct from built-in `github-copilot` device login (bundled, disabled by default)
Clawdbot plugins are **TypeScript modules** loaded at runtime via jiti. They can
register:
@ -56,6 +59,7 @@ register:
- Optional config validation
Plugins run **inprocess** with the Gateway, so treat them as trusted code.
Tool authoring guide: [Plugin agent tools](/plugins/agent-tools).
## Discovery & precedence
@ -136,6 +140,24 @@ Fields:
Config changes **require a gateway restart**.
## Plugin slots (exclusive categories)
Some plugin categories are **exclusive** (only one active at a time). Use
`plugins.slots` to select which plugin owns the slot:
```json5
{
plugins: {
slots: {
memory: "memory-core" // or "none" to disable memory plugins
}
}
}
```
If multiple plugins declare `kind: "memory"`, only the selected one loads. Others
are disabled with diagnostics.
## Control UI (schema + labels)
The Control UI uses `config.schema` (JSON Schema + `uiHints`) to render better forms.
@ -194,6 +216,27 @@ Plugins export either:
- A function: `(api) => { ... }`
- An object: `{ id, name, configSchema, register(api) { ... } }`
## Plugin hooks
Plugins can ship hooks and register them at runtime. This lets a plugin bundle
event-driven automation without a separate hook pack install.
### Example
```
import { registerPluginHooksFromDir } from "clawdbot/plugin-sdk";
export default function register(api) {
registerPluginHooksFromDir(api, "./hooks");
}
```
Notes:
- Hook directories follow the normal hook structure (`HOOK.md` + `handler.ts`).
- Hook eligibility rules still apply (OS/bins/env/config requirements).
- Plugin-managed hooks show up in `clawdbot hooks list` with `plugin:<id>`.
- You cannot enable/disable plugin-managed hooks via `clawdbot hooks`; enable/disable the plugin instead.
## Provider plugins (model auth)
Plugins can register **model provider auth** flows so users can run OAuth or
@ -359,24 +402,9 @@ export default function (api) {
Load the plugin (extensions dir or `plugins.load.paths`), restart the gateway,
then configure `channels.<id>` in your config.
### Register a tool
### Agent tools
```ts
import { Type } from "@sinclair/typebox";
export default function (api) {
api.registerTool({
name: "my_tool",
description: "Do a thing",
parameters: Type.Object({
input: Type.String(),
}),
async execute(_id, params) {
return { content: [{ type: "text", text: params.input }] };
},
});
}
```
See the dedicated guide: [Plugin agent tools](/plugins/agent-tools).
### Register a gateway RPC method

View File

@ -0,0 +1,94 @@
---
summary: "Write agent tools in a plugin (schemas, optional tools, allowlists)"
read_when:
- You want to add a new agent tool in a plugin
- You need to make a tool opt-in via allowlists
---
# Plugin agent tools
Clawdbot plugins can register **agent tools** (JSONschema functions) that are exposed
to the LLM during agent runs. Tools can be **required** (always available) or
**optional** (optin).
Agent tools are configured under `tools` in the main config, or peragent under
`agents.list[].tools`. The allowlist/denylist policy controls which tools the agent
can call.
## Basic tool
```ts
import { Type } from "@sinclair/typebox";
export default function (api) {
api.registerTool({
name: "my_tool",
description: "Do a thing",
parameters: Type.Object({
input: Type.String(),
}),
async execute(_id, params) {
return { content: [{ type: "text", text: params.input }] };
},
});
}
```
## Optional tool (optin)
Optional tools are **never** autoenabled. Users must add them to an agent
allowlist.
```ts
export default function (api) {
api.registerTool(
{
name: "workflow_tool",
description: "Run a local workflow",
parameters: {
type: "object",
properties: {
pipeline: { type: "string" },
},
required: ["pipeline"],
},
async execute(_id, params) {
return { content: [{ type: "text", text: params.pipeline }] };
},
},
{ optional: true },
);
}
```
Enable optional tools in `agents.list[].tools.allow` (or global `tools.allow`):
```json5
{
agents: {
list: [
{
id: "main",
tools: {
allow: [
"workflow_tool", // specific tool name
"workflow", // plugin id (enables all tools from that plugin)
"group:plugins" // all plugin tools
]
}
}
]
}
}
```
Other config knobs that affect tool availability:
- `tools.profile` / `agents.list[].tools.profile` (base allowlist)
- `tools.byProvider` / `agents.list[].tools.byProvider` (providerspecific allow/deny)
- `tools.sandbox.tools.*` (sandbox tool policy when sandboxed)
## Rules + tips
- Tool names must **not** clash with core tool names; conflicting tools are skipped.
- Plugin ids used in allowlists must not clash with core tool names.
- Prefer `optional: true` for tools that trigger side effects or require extra
binaries/credentials.

View File

@ -0,0 +1,70 @@
---
summary: "Sign in to GitHub Copilot from Clawdbot using the device flow"
read_when:
- You want to use GitHub Copilot as a model provider
- You need the `clawdbot models auth login-github-copilot` flow
---
# Github Copilot
## What is GitHub Copilot?
GitHub Copilot is GitHub's AI coding assistant. It provides access to Copilot
models for your GitHub account and plan. Clawdbot can use Copilot as a model
provider in two different ways.
## Two ways to use Copilot in Clawdbot
### 1) Built-in GitHub Copilot provider (`github-copilot`)
Use the native device-login flow to obtain a GitHub token, then exchange it for
Copilot API tokens when Clawdbot runs. This is the **default** and simplest path
because it does not require VS Code.
### 2) Copilot Proxy plugin (`copilot-proxy`)
Use the **Copilot Proxy** VS Code extension as a local bridge. Clawdbot talks to
the proxys `/v1` endpoint and uses the model list you configure there. Choose
this when you already run Copilot Proxy in VS Code or need to route through it.
You must enable the plugin and keep the VS Code extension running.
Use GitHub Copilot as a model provider (`github-copilot`). The login command runs
the GitHub device flow, saves an auth profile, and updates your config to use that
profile.
## CLI setup
```bash
clawdbot models auth login-github-copilot
```
You'll be prompted to visit a URL and enter a one-time code. Keep the terminal
open until it completes.
### Optional flags
```bash
clawdbot models auth login-github-copilot --profile-id github-copilot:work
clawdbot models auth login-github-copilot --yes
```
## Set a default model
```bash
clawdbot models set github-copilot/gpt-4o
```
### Config snippet
```json5
{
agents: { defaults: { model: { primary: "github-copilot/gpt-4o" } } }
}
```
## Notes
- Requires an interactive TTY; run it directly in a terminal.
- Copilot model availability depends on your plan; if a model is rejected, try
another ID (for example `github-copilot/gpt-4.1`).
- The login stores a GitHub token in the auth profile store and exchanges it for a
Copilot API token when Clawdbot runs.

View File

@ -26,6 +26,7 @@ Looking for chat channel docs (WhatsApp/Telegram/Discord/Slack/etc.)? See [Chann
- [OpenAI (API + Codex)](/providers/openai)
- [Anthropic (API + Claude Code CLI)](/providers/anthropic)
- [Qwen (OAuth)](/providers/qwen)
- [OpenRouter](/providers/openrouter)
- [Vercel AI Gateway](/providers/vercel-ai-gateway)
- [Moonshot AI (Kimi + Kimi Code)](/providers/moonshot)

51
docs/providers/qwen.md Normal file
View File

@ -0,0 +1,51 @@
---
summary: "Use Qwen OAuth (free tier) in Clawdbot"
read_when:
- You want to use Qwen with Clawdbot
- You want free-tier OAuth access to Qwen Coder
---
# Qwen
Qwen provides a free-tier OAuth flow for Qwen Coder and Qwen Vision models
(2,000 requests/day, subject to Qwen rate limits).
## Enable the plugin
```bash
clawdbot plugins enable qwen-portal-auth
```
Restart the Gateway after enabling.
## Authenticate
```bash
clawdbot models auth login --provider qwen-portal --set-default
```
This runs the Qwen device-code OAuth flow and writes a provider entry to your
`models.json` (plus a `qwen` alias for quick switching).
## Model IDs
- `qwen-portal/coder-model`
- `qwen-portal/vision-model`
Switch models with:
```bash
clawdbot models set qwen-portal/coder-model
```
## Reuse Qwen Code CLI login
If you already logged in with the Qwen Code CLI, Clawdbot will sync credentials
from `~/.qwen/oauth_creds.json` when it loads the auth store. You still need a
`models.providers.qwen-portal` entry (use the login command above to create one).
## Notes
- Tokens auto-refresh; re-run the login command if refresh fails or access is revoked.
- Default base URL: `https://portal.qwen.ai/v1` (override with
`models.providers.qwen-portal.baseUrl` if Qwen provides a different endpoint).
- See [Model providers](/concepts/model-providers) for provider-wide rules.

265
docs/refactor/exec-host.md Normal file
View File

@ -0,0 +1,265 @@
---
summary: "Refactor plan: exec host routing, node approvals, and headless runner"
read_when:
- Designing exec host routing or exec approvals
- Implementing node runner + UI IPC
- Adding exec host security modes and slash commands
---
# Exec host refactor plan
## Goals
- Add `exec.host` + `exec.security` to route execution across **sandbox**, **gateway**, and **node**.
- Keep defaults **safe**: no cross-host execution unless explicitly enabled.
- Split execution into a **headless runner service** with optional UI (macOS app) via local IPC.
- Provide **per-agent** policy, allowlist, ask mode, and node binding.
- Support **ask modes** that work *with* or *without* allowlists.
- Cross-platform: Unix socket + token auth (macOS/Linux/Windows parity).
## Non-goals
- No legacy allowlist migration or legacy schema support.
- No PTY/streaming for node exec (aggregated output only).
- No new network layer beyond the existing Bridge + Gateway.
## Decisions (locked)
- **Config keys:** `exec.host` + `exec.security` (per-agent override allowed).
- **Elevation:** keep `/elevated` as an alias for gateway full access.
- **Ask default:** `on-miss`.
- **Approvals store:** `~/.clawdbot/exec-approvals.json` (JSON, no legacy migration).
- **Runner:** headless system service; UI app hosts a Unix socket for approvals.
- **Node identity:** use existing `nodeId`.
- **Socket auth:** Unix socket + token (cross-platform); split later if needed.
- **Node host state:** `~/.clawdbot/node.json` (node id + pairing token).
- **macOS exec host:** run `system.run` inside the macOS app; node service forwards requests over local IPC.
- **No XPC helper:** stick to Unix socket + token + peer checks.
## Key concepts
### Host
- `sandbox`: Docker exec (current behavior).
- `gateway`: exec on gateway host.
- `node`: exec on node runner via Bridge (`system.run`).
### Security mode
- `deny`: always block.
- `allowlist`: allow only matches.
- `full`: allow everything (equivalent to elevated).
### Ask mode
- `off`: never ask.
- `on-miss`: ask only when allowlist does not match.
- `always`: ask every time.
Ask is **independent** of allowlist; allowlist can be used with `always` or `on-miss`.
### Policy resolution (per exec)
1) Resolve `exec.host` (tool param → agent override → global default).
2) Resolve `exec.security` and `exec.ask` (same precedence).
3) If host is `sandbox`, proceed with local sandbox exec.
4) If host is `gateway` or `node`, apply security + ask policy on that host.
## Default safety
- Default `exec.host = sandbox`.
- Default `exec.security = deny` for `gateway` and `node`.
- Default `exec.ask = on-miss` (only relevant if security allows).
- If no node binding is set, **agent may target any node**, but only if policy allows it.
## Config surface
### Tool parameters
- `exec.host` (optional): `sandbox | gateway | node`.
- `exec.security` (optional): `deny | allowlist | full`.
- `exec.ask` (optional): `off | on-miss | always`.
- `exec.node` (optional): node id/name to use when `host=node`.
### Config keys (global)
- `tools.exec.host`
- `tools.exec.security`
- `tools.exec.ask`
- `tools.exec.node` (default node binding)
### Config keys (per agent)
- `agents.list[].tools.exec.host`
- `agents.list[].tools.exec.security`
- `agents.list[].tools.exec.ask`
- `agents.list[].tools.exec.node`
### Alias
- `/elevated on` = set `tools.exec.host=gateway`, `tools.exec.security=full` for the agent session.
- `/elevated off` = restore previous exec settings for the agent session.
## Approvals store (JSON)
Path: `~/.clawdbot/exec-approvals.json`
Purpose:
- Local policy + allowlists for the **execution host** (gateway or node runner).
- Ask fallback when no UI is available.
- IPC credentials for UI clients.
Proposed schema (v1):
```json
{
"version": 1,
"socket": {
"path": "~/.clawdbot/exec-approvals.sock",
"token": "base64-opaque-token"
},
"defaults": {
"security": "deny",
"ask": "on-miss",
"askFallback": "deny"
},
"agents": {
"agent-id-1": {
"security": "allowlist",
"ask": "on-miss",
"allowlist": [
{
"pattern": "~/Projects/**/bin/rg",
"lastUsedAt": 0,
"lastUsedCommand": "rg -n TODO",
"lastResolvedPath": "/Users/user/Projects/.../bin/rg"
}
]
}
}
}
```
Notes:
- No legacy allowlist formats.
- `askFallback` applies only when `ask` is required and no UI is reachable.
- File permissions: `0600`.
## Runner service (headless)
### Role
- Enforce `exec.security` + `exec.ask` locally.
- Execute system commands and return output.
- Emit Bridge events for exec lifecycle (optional but recommended).
### Service lifecycle
- Launchd/daemon on macOS; system service on Linux/Windows.
- Approvals JSON is local to the execution host.
- UI hosts a local Unix socket; runners connect on demand.
## UI integration (macOS app)
### IPC
- Unix socket at `~/.clawdbot/exec-approvals.sock` (0600).
- Token stored in `exec-approvals.json` (0600).
- Peer checks: same-UID only.
- Challenge/response: nonce + HMAC(token, request-hash) to prevent replay.
- Short TTL (e.g., 10s) + max payload + rate limit.
### Ask flow (macOS app exec host)
1) Node service receives `system.run` from gateway.
2) Node service connects to the local socket and sends the prompt/exec request.
3) App validates peer + token + HMAC + TTL, then shows dialog if needed.
4) App executes the command in UI context and returns output.
5) Node service returns output to gateway.
If UI missing:
- Apply `askFallback` (`deny|allowlist|full`).
### Diagram (SCI)
```
Agent -> Gateway -> Bridge -> Node Service (TS)
| IPC (UDS + token + HMAC + TTL)
v
Mac App (UI + TCC + system.run)
```
## Node identity + binding
- Use existing `nodeId` from Bridge pairing.
- Binding model:
- `tools.exec.node` restricts the agent to a specific node.
- If unset, agent can pick any node (policy still enforces defaults).
- Node selection resolution:
- `nodeId` exact match
- `displayName` (normalized)
- `remoteIp`
- `nodeId` prefix (>= 6 chars)
## Eventing
### Who sees events
- System events are **per session** and shown to the agent on the next prompt.
- Stored in the gateway in-memory queue (`enqueueSystemEvent`).
### Event text
- `Exec started (node=<id>, id=<runId>)`
- `Exec finished (node=<id>, id=<runId>, code=<code>)` + optional output tail
- `Exec denied (node=<id>, id=<runId>, <reason>)`
### Transport
Option A (recommended):
- Runner sends Bridge `event` frames `exec.started` / `exec.finished`.
- Gateway `handleBridgeEvent` maps these into `enqueueSystemEvent`.
Option B:
- Gateway `exec` tool handles lifecycle directly (synchronous only).
## Exec flows
### Sandbox host
- Existing `exec` behavior (Docker or host when unsandboxed).
- PTY supported in non-sandbox mode only.
### Gateway host
- Gateway process executes on its own machine.
- Enforces local `exec-approvals.json` (security/ask/allowlist).
### Node host
- Gateway calls `node.invoke` with `system.run`.
- Runner enforces local approvals.
- Runner returns aggregated stdout/stderr.
- Optional Bridge events for start/finish/deny.
## Output caps
- Cap combined stdout+stderr at **200k**; keep **tail 20k** for events.
- Truncate with a clear suffix (e.g., `"… (truncated)"`).
## Slash commands
- `/exec host=<sandbox|gateway|node> security=<deny|allowlist|full> ask=<off|on-miss|always> node=<id>`
- Per-agent, per-session overrides; non-persistent unless saved via config.
- `/elevated on|off` remains a shortcut for `host=gateway security=full`.
## Cross-platform story
- The runner service is the portable execution target.
- UI is optional; if missing, `askFallback` applies.
- Windows/Linux support the same approvals JSON + socket protocol.
## Implementation phases
### Phase 1: config + exec routing
- Add config schema for `exec.host`, `exec.security`, `exec.ask`, `exec.node`.
- Update tool plumbing to respect `exec.host`.
- Add `/exec` slash command and keep `/elevated` alias.
### Phase 2: approvals store + gateway enforcement
- Implement `exec-approvals.json` reader/writer.
- Enforce allowlist + ask modes for `gateway` host.
- Add output caps.
### Phase 3: node runner enforcement
- Update node runner to enforce allowlist + ask.
- Add Unix socket prompt bridge to macOS app UI.
- Wire `askFallback`.
### Phase 4: events
- Add node → gateway Bridge events for exec lifecycle.
- Map to `enqueueSystemEvent` for agent prompts.
### Phase 5: UI polish
- Mac app: allowlist editor, per-agent switcher, ask policy UI.
- Node binding controls (optional).
## Testing plan
- Unit tests: allowlist matching (glob + case-insensitive).
- Unit tests: policy resolution precedence (tool param → agent override → global).
- Integration tests: node runner deny/allow/ask flows.
- Bridge event tests: node event → system event routing.
## Open risks
- UI unavailability: ensure `askFallback` is respected.
- Long-running commands: rely on timeout + output caps.
- Multi-node ambiguity: error unless node binding or explicit node param.
## Related docs
- [Exec tool](/tools/exec)
- [Exec approvals](/tools/exec-approvals)
- [Nodes](/nodes)
- [Elevated mode](/tools/elevated)

187
docs/refactor/plugin-sdk.md Normal file
View File

@ -0,0 +1,187 @@
---
summary: "Plan: one clean plugin SDK + runtime for all messaging connectors"
read_when:
- Defining or refactoring the plugin architecture
- Migrating channel connectors to the plugin SDK/runtime
---
# Plugin SDK + Runtime Refactor Plan
Goal: every messaging connector is a plugin (bundled or external) using one stable API.
No plugin imports from `src/**` directly. All dependencies go through the SDK or runtime.
## Why now
- Current connectors mix patterns: direct core imports, dist-only bridges, and custom helpers.
- This makes upgrades brittle and blocks a clean external plugin surface.
## Target architecture (two layers)
### 1) Plugin SDK (compile-time, stable, publishable)
Scope: types, helpers, and config utilities. No runtime state, no side effects.
Contents (examples):
- Types: `ChannelPlugin`, adapters, `ChannelMeta`, `ChannelCapabilities`, `ChannelDirectoryEntry`.
- Config helpers: `buildChannelConfigSchema`, `setAccountEnabledInConfigSection`, `deleteAccountFromConfigSection`,
`applyAccountNameToChannelSection`.
- Pairing helpers: `PAIRING_APPROVED_MESSAGE`, `formatPairingApproveHint`.
- Onboarding helpers: `promptChannelAccessConfig`, `addWildcardAllowFrom`, onboarding types.
- Tool param helpers: `createActionGate`, `readStringParam`, `readNumberParam`, `readReactionParams`, `jsonResult`.
- Docs link helper: `formatDocsLink`.
Delivery:
- Publish as `@clawdbot/plugin-sdk` (or export from core under `clawdbot/plugin-sdk`).
- Semver with explicit stability guarantees.
### 2) Plugin Runtime (execution surface, injected)
Scope: everything that touches core runtime behavior.
Accessed via `ClawdbotPluginApi.runtime` so plugins never import `src/**`.
Proposed surface (minimal but complete):
```ts
export type PluginRuntime = {
channel: {
text: {
chunkMarkdownText(text: string, limit: number): string[];
resolveTextChunkLimit(cfg: ClawdbotConfig, channel: string, accountId?: string): number;
hasControlCommand(text: string, cfg: ClawdbotConfig): boolean;
};
reply: {
dispatchReplyWithBufferedBlockDispatcher(params: {
ctx: unknown;
cfg: unknown;
dispatcherOptions: {
deliver: (payload: { text?: string; mediaUrls?: string[]; mediaUrl?: string }) =>
void | Promise<void>;
onError?: (err: unknown, info: { kind: string }) => void;
};
}): Promise<void>;
createReplyDispatcherWithTyping?: unknown; // adapter for Teams-style flows
};
routing: {
resolveAgentRoute(params: {
cfg: unknown;
channel: string;
accountId: string;
peer: { kind: "dm" | "group" | "channel"; id: string };
}): { sessionKey: string; accountId: string };
};
pairing: {
buildPairingReply(params: { channel: string; idLine: string; code: string }): string;
readAllowFromStore(channel: string): Promise<string[]>;
upsertPairingRequest(params: {
channel: string;
id: string;
meta?: { name?: string };
}): Promise<{ code: string; created: boolean }>;
};
media: {
fetchRemoteMedia(params: { url: string }): Promise<{ buffer: Buffer; contentType?: string }>;
saveMediaBuffer(
buffer: Uint8Array,
contentType: string | undefined,
direction: "inbound" | "outbound",
maxBytes: number,
): Promise<{ path: string; contentType?: string }>;
};
mentions: {
buildMentionRegexes(cfg: ClawdbotConfig, agentId?: string): RegExp[];
matchesMentionPatterns(text: string, regexes: RegExp[]): boolean;
};
groups: {
resolveGroupPolicy(cfg: ClawdbotConfig, channel: string, accountId: string, groupId: string): {
allowlistEnabled: boolean;
allowed: boolean;
groupConfig?: unknown;
defaultConfig?: unknown;
};
resolveRequireMention(
cfg: ClawdbotConfig,
channel: string,
accountId: string,
groupId: string,
override?: boolean,
): boolean;
};
debounce: {
createInboundDebouncer<T>(opts: {
debounceMs: number;
buildKey: (v: T) => string | null;
shouldDebounce: (v: T) => boolean;
onFlush: (entries: T[]) => Promise<void>;
onError?: (err: unknown) => void;
}): { push: (v: T) => void; flush: () => Promise<void> };
resolveInboundDebounceMs(cfg: ClawdbotConfig, channel: string): number;
};
commands: {
resolveCommandAuthorizedFromAuthorizers(params: {
useAccessGroups: boolean;
authorizers: Array<{ configured: boolean; allowed: boolean }>;
}): boolean;
};
};
logging: {
shouldLogVerbose(): boolean;
getChildLogger(name: string): PluginLogger;
};
state: {
resolveStateDir(cfg: ClawdbotConfig): string;
};
};
```
Notes:
- Runtime is the only way to access core behavior.
- SDK is intentionally small and stable.
- Each runtime method maps to an existing core implementation (no duplication).
## Migration plan (phased, safe)
### Phase 0: scaffolding
- Introduce `@clawdbot/plugin-sdk`.
- Add `api.runtime` to `ClawdbotPluginApi` with the surface above.
- Maintain existing imports during a transition window (deprecation warnings).
### Phase 1: bridge cleanup (low risk)
- Replace per-extension `core-bridge.ts` with `api.runtime`.
- Migrate BlueBubbles, Zalo, Zalo Personal first (already close).
- Remove duplicated bridge code.
### Phase 2: light direct-import plugins
- Migrate Matrix to SDK + runtime.
- Validate onboarding, directory, group mention logic.
### Phase 3: heavy direct-import plugins
- Migrate MS Teams (largest set of runtime helpers).
- Ensure reply/typing semantics match current behavior.
### Phase 4: iMessage pluginization
- Move iMessage into `extensions/imessage`.
- Replace direct core calls with `api.runtime`.
- Keep config keys, CLI behavior, and docs intact.
### Phase 5: enforcement
- Add lint rule / CI check: no `extensions/**` imports from `src/**`.
- Add plugin SDK/version compatibility checks (runtime + SDK semver).
## Compatibility and versioning
- SDK: semver, published, documented changes.
- Runtime: versioned per core release. Add `api.runtime.version`.
- Plugins declare a required runtime range (e.g., `clawdbotRuntime: ">=2026.2.0"`).
## Testing strategy
- Adapter-level unit tests (runtime functions exercised with real core implementation).
- Golden tests per plugin: ensure no behavior drift (routing, pairing, allowlist, mention gating).
- A single end-to-end plugin sample used in CI (install + run + smoke).
## Open questions
- Where to host SDK types: separate package or core export?
- Runtime type distribution: in SDK (types only) or in core?
- How to expose docs links for bundled vs external plugins?
- Do we allow limited direct core imports for in-repo plugins during transition?
## Success criteria
- All channel connectors are plugins using SDK + runtime.
- No `extensions/**` imports from `src/**`.
- New connector templates depend only on SDK + runtime.
- External plugins can be developed and updated without core source access.
Related docs: [Plugins](/plugin), [Channels](/channels/index), [Configuration](/gateway/configuration).

View File

@ -0,0 +1,116 @@
---
summary: "Audit what can spend money, which keys are used, and how to view usage"
read_when:
- You want to understand which features may call paid APIs
- You need to audit keys, costs, and usage visibility
- Youre explaining /status or /usage cost reporting
---
# API usage & costs
This doc lists **features that can invoke API keys** and where their costs show up. It focuses on
Clawdbot features that can generate provider usage or paid API calls.
## Where costs show up (chat + CLI)
**Per-session cost snapshot**
- `/status` shows the current session model, context usage, and last response tokens.
- If the model uses **API-key auth**, `/status` also shows **estimated cost** for the last reply.
**Per-message cost footer**
- `/usage full` appends a usage footer to every reply, including **estimated cost** (API-key only).
- `/usage tokens` shows tokens only; OAuth flows hide dollar cost.
**CLI usage windows (provider quotas)**
- `clawdbot status --usage` and `clawdbot channels list` show provider **usage windows**
(quota snapshots, not per-message costs).
See [Token use & costs](/token-use) for details and examples.
## How keys are discovered
Clawdbot can pick up credentials from:
- **Auth profiles** (per-agent, stored in `auth-profiles.json`).
- **Environment variables** (e.g. `OPENAI_API_KEY`, `BRAVE_API_KEY`, `FIRECRAWL_API_KEY`).
- **Config** (`models.providers.*.apiKey`, `tools.web.search.*`, `tools.web.fetch.firecrawl.*`,
`memorySearch.*`, `talk.apiKey`).
- **Skills** (`skills.entries.<name>.apiKey`) which may export keys to the skill process env.
## Features that can spend keys
### 1) Core model responses (chat + tools)
Every reply or tool call uses the **current model provider** (OpenAI, Anthropic, etc). This is the
primary source of usage and cost.
See [Models](/providers/models) for pricing config and [Token use & costs](/token-use) for display.
### 2) Media understanding (audio/image/video)
Inbound media can be summarized/transcribed before the reply runs. This uses model/provider APIs.
- Audio: OpenAI / Groq / Deepgram (now **auto-enabled** when keys exist).
- Image: OpenAI / Anthropic / Google.
- Video: Google.
See [Media understanding](/nodes/media-understanding).
### 3) Memory embeddings + semantic search
Semantic memory search uses **embedding APIs** when configured for remote providers:
- `memorySearch.provider = "openai"` → OpenAI embeddings
- `memorySearch.provider = "gemini"` → Gemini embeddings
- Optional fallback to OpenAI if local embeddings fail
You can keep it local with `memorySearch.provider = "local"` (no API usage).
See [Memory](/concepts/memory).
### 4) Web search tool (Brave / Perplexity via OpenRouter)
`web_search` uses API keys and may incur usage charges:
- **Brave Search API**: `BRAVE_API_KEY` or `tools.web.search.apiKey`
- **Perplexity** (via OpenRouter): `PERPLEXITY_API_KEY` or `OPENROUTER_API_KEY`
**Brave free tier (generous):**
- **2,000 requests/month**
- **1 request/second**
- **Credit card required** for verification (no charge unless you upgrade)
See [Web tools](/tools/web).
### 5) Web fetch tool (Firecrawl)
`web_fetch` can call **Firecrawl** when an API key is present:
- `FIRECRAWL_API_KEY` or `tools.web.fetch.firecrawl.apiKey`
If Firecrawl isnt configured, the tool falls back to direct fetch + readability (no paid API).
See [Web tools](/tools/web).
### 6) Provider usage snapshots (status/health)
Some status commands call **provider usage endpoints** to display quota windows or auth health.
These are typically low-volume calls but still hit provider APIs:
- `clawdbot status --usage`
- `clawdbot models status --json`
See [Models CLI](/cli/models).
### 7) Compaction safeguard summarization
The compaction safeguard can summarize session history using the **current model**, which
invokes provider APIs when it runs.
See [Session management + compaction](/reference/session-management-compaction).
### 8) Model scan / probe
`clawdbot models scan` can probe OpenRouter models and uses `OPENROUTER_API_KEY` when
probing is enabled.
See [Models CLI](/cli/models).
### 9) Talk (speech)
Talk mode can invoke **ElevenLabs** when configured:
- `ELEVENLABS_API_KEY` or `talk.apiKey`
See [Talk mode](/nodes/talk).
### 10) Skills (third-party APIs)
Skills can store `apiKey` in `skills.entries.<name>.apiKey`. If a skill uses that key for external
APIs, it can incur costs according to the skills provider.
See [Skills](/tools/skills).

View File

@ -82,7 +82,8 @@ Each `sessionKey` points at a current `sessionId` (the transcript file that cont
Rules of thumb:
- **Reset** (`/new`, `/reset`) creates a new `sessionId` for that `sessionKey`.
- **Idle expiry** (`session.idleMinutes`) creates a new `sessionId` when a message arrives after the idle window.
- **Daily reset** (default 4:00 AM local time on the gateway host) creates a new `sessionId` on the next message after the reset boundary.
- **Idle expiry** (`session.reset.idleMinutes` or legacy `session.idleMinutes`) creates a new `sessionId` when a message arrives after the idle window. When daily + idle are both configured, whichever expires first wins.
Implementation detail: the decision happens in `initSessionState()` in `src/auto-reply/reply/session.ts`.

Some files were not shown because too many files have changed in this diff Show More