openclaw/CLAWDBOT-SECURITY-PLAN.md
ronitchidara 2fd698d137 security: add security documentation and doctor credential check
Security Documentation:
- docs/security/threat-model.md: Comprehensive threat model covering
  channels, tools, gateway, local files, browser automation, and plugins
- docs/security/data-handling.md: Data handling policy covering storage,
  retention, user rights (access, export, delete), and compliance

Doctor Security Check:
- Added credential encryption check to doctor-security.ts
- Warns about unencrypted WhatsApp/Web credentials

This completes Phase 1 Security Hardening (Tasks 1.1-1.6).

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
2026-01-28 16:55:18 +05:30

9.0 KiB

Clawdbot Security Hardening Plan - Phase 1

Location: This file lives in the security worktree at ~/clawdbot-security Branch: phase1-security Purpose: Tracks progress on Phase 1 Security Hardening tasks. Other Claude Code sessions can pick up incomplete tasks.


Worktree Setup

Git worktrees are set up for parallel development:

~/clawdbot-analysis/        # main branch (original clone)
~/clawdbot-security/        # phase1-security branch (THIS WORKTREE)

To create additional worktrees for other phases:

cd ~/clawdbot-analysis
git worktree add ../clawdbot-architecture -b phase2-architecture
git worktree add ../clawdbot-ux -b phase3-ux

Phase 1 Task Status

COMPLETED

1.1 Prompt Injection Defense

  • Files Modified:
    • src/gateway/chat-sanitize.ts - Added 300+ lines of injection detection
    • src/gateway/chat-sanitize.test.ts - 38 tests
  • Features:
    • Regex-based jailbreak detection with tiered severity (critical/high/medium/low)
    • Prompt boundary markers ([USER_INPUT_START]/[USER_INPUT_END])
    • Configurable blocking, logging, and warning features
    • Patterns: ignore instructions, DAN/jailbreak, system prompt extraction, role manipulation
  • Commit: 77c93a8dc "security: add prompt injection defense system"

1.2 Command Execution Blocklist

  • Files Created/Modified:
    • src/infra/exec-blocklist.ts - New blocklist module (~400 lines)
    • src/infra/exec-blocklist.test.ts - 47 tests
    • src/infra/exec-approvals.ts - Integrated blocklist check
    • src/infra/exec-approvals.test.ts - Updated for blocklist behavior
  • Features:
    • Critical: rm -rf /, dd to disk, mkfs, halt/reboot, fork bombs (ALWAYS BLOCKED)
    • High: sudo, passwd, visudo, iptables, user management (BLOCKED, requires explicit approval)
    • Medium: command substitution, eval, curl POST (WARNED, allowed by default)
    • Blocklist is checked BEFORE allowlist evaluation
  • Commit: 33eeb033c "security: add command execution blocklist"

1.3 Secrets Manager (Core + Integration)

  • Files Created/Modified:
    • src/infra/secrets-manager.ts - Cross-platform secrets manager (~700 lines)
    • src/infra/secrets-manager.test.ts - 27 tests
    • src/discord/token.ts - Added secrets as Priority 1 source
    • src/discord/token.test.ts - Updated tests with skipSecrets option
    • src/discord/accounts.ts - Updated type to include "secrets" source
    • src/telegram/token.ts - Added secrets as Priority 1 source
    • src/telegram/token.test.ts - Updated tests with skipSecrets option
    • src/telegram/accounts.ts - Updated type to include "secrets" source
    • src/slack/accounts.ts - Added secrets as Priority 1 source with skipSecrets option
    • src/web/auth-store.ts - Added encrypted credentials support for WhatsApp
  • Features:
    • macOS Keychain via security command
    • Linux Secret Service via secret-tool (libsecret)
    • Fallback to AES-256-GCM encrypted files with PBKDF2 key derivation
    • Machine-derived encryption key for file fallback
    • Utilities: generateSecureToken, hashSecret, verifySecretHash
    • File encryption utilities: encryptFile, decryptFile, readPossiblyEncryptedFile
    • WhatsApp creds.json now supports .enc encrypted format
    • Migration helper: migrateWebCredsToEncrypted() for WhatsApp
    • All token resolution functions now check secrets-manager first
  • Commits:
    • a6f48ac29 "security: add cross-platform secrets manager"
    • TBD "security: integrate secrets-manager with all token stores"
  • TODO:
    • Add migration CLI command: clawdbot secrets migrate
    • Update clawdbot doctor to check for unencrypted credentials

1.4 Gateway Authentication & Rate Limiting

  • Files Created/Modified:
    • src/gateway/rate-limit.ts - New rate limiting module (~370 lines)
    • src/gateway/rate-limit.test.ts - 19 tests
    • src/gateway/server-startup-log.ts - Added security warnings
    • src/gateway/server-startup-log.test.ts - 9 tests
    • src/gateway/server.impl.ts - Integrated rate limiter and security warnings
    • src/config/types.gateway.ts - Added GatewayRateLimitConfig type
  • Features:
    • Token bucket algorithm for smooth rate limiting with burst support
    • Per-client tracking (separate buckets per IP/session)
    • Configurable limits: unauthenticated (60/min), authenticated (unlimited by default)
    • Channel message rate limiting (200/min per channel)
    • Exponential backoff after auth failures (1s base, 60s max)
    • Security warning at startup if binding to non-loopback without auth
    • All settings configurable in gateway.rateLimit config
  • Commit: TBD

1.5 Pairing & Approval Hardening

  • Files Modified:
    • src/pairing/pairing-store.ts - Pairing security hardening (~120 lines added)
    • src/pairing/pairing-store.test.ts - 11 tests (6 new security tests)
    • src/infra/exec-approvals.ts - One-time-use nonce system (~100 lines added)
    • src/infra/exec-approvals.test.ts - 38 tests (8 new nonce tests)
  • Features:
    • Pairing codes increased from 8 chars (40 bits) to 16 chars (80 bits entropy)
    • HMAC-SHA256 signatures on pairing stores for integrity verification
    • Rate limiting on pairing approval attempts (10/minute per channel)
    • Automatic store reset on signature verification failure (tampering detection)
    • One-time-use nonces for exec approval requests (replay protection)
    • Nonce TTL (5 minutes) with automatic cleanup
    • Request ID verification on approval responses
    • All nonce functions exported for testing: generateApprovalNonce, verifyAndConsumeNonce, isNonceValid
  • Commit: TBD

1.6 Security Documentation

  • Files Created:
    • docs/security/threat-model.md - Comprehensive threat model (~250 lines)
    • docs/security/data-handling.md - Data handling policy (~280 lines)
  • Files Modified:
    • src/commands/doctor-security.ts - Added credential encryption check
  • Features:
    • Threat model covers: channels, tools, gateway, local files, browser automation, plugins
    • Documents threat actors, attack surfaces, mitigations, residual risks
    • Data handling policy covers: storage, retention, user rights, compliance
    • Doctor now warns about unencrypted WhatsApp/Web credentials
  • Commit: TBD

ALL PHASE 1 TASKS COMPLETE


How to Continue This Work

For a New Claude Code Session:

  1. Navigate to the worktree:

    cd ~/clawdbot-security
    
  2. Read this plan file:

    cat CLAWDBOT-SECURITY-PLAN.md
    
  3. Check current branch and status:

    git log --oneline -5
    git status
    
  4. Pick an IN PROGRESS or PENDING task and continue.

  5. Follow commit conventions:

    git commit -m "security: <description>
    
    Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>"
    

Parallel Work Guidelines:

  • Each task modifies different files, so parallel work is safe
  • If touching the same file, coordinate or work sequentially
  • Run npx vitest run <path> after changes to verify tests pass

Testing Commands

# Run specific test file
npx vitest run src/gateway/chat-sanitize.test.ts
npx vitest run src/infra/exec-blocklist.test.ts
npx vitest run src/infra/secrets-manager.test.ts

# Run all tests
pnpm test

# Type check
pnpm build

Files Changed Summary

Phase File Status Lines Changed
1.1 src/gateway/chat-sanitize.ts +320
1.1 src/gateway/chat-sanitize.test.ts +200
1.2 src/infra/exec-blocklist.ts +450
1.2 src/infra/exec-blocklist.test.ts +250
1.2 src/infra/exec-approvals.ts +50
1.3 src/infra/secrets-manager.ts +700
1.3 src/infra/secrets-manager.test.ts +270
1.3 src/discord/token.ts +25
1.3 src/discord/token.test.ts +10
1.3 src/discord/accounts.ts +5
1.3 src/telegram/token.ts +30
1.3 src/telegram/token.test.ts +10
1.3 src/telegram/accounts.ts +5
1.3 src/slack/accounts.ts +40
1.3 src/web/auth-store.ts +120
1.4 src/gateway/rate-limit.ts +370
1.4 src/gateway/rate-limit.test.ts +310
1.4 src/gateway/server-startup-log.ts +60
1.4 src/gateway/server-startup-log.test.ts +170
1.4 src/gateway/server.impl.ts +15
1.4 src/config/types.gateway.ts +45
1.5 src/pairing/pairing-store.ts +120
1.5 src/pairing/pairing-store.test.ts +80
1.5 src/infra/exec-approvals.ts +100
1.5 src/infra/exec-approvals.test.ts +70
1.6 docs/security/threat-model.md +250
1.6 docs/security/data-handling.md +280
1.6 src/commands/doctor-security.ts +30

Contact / Context

  • Original Plan: /Users/ronitchidara/Desktop/clawdbot-production-readiness-analysis.md
  • Main Codebase: ~/clawdbot-analysis
  • Security Worktree: ~/clawdbot-security (this location)

Last updated: 2026-01-27 by Claude Opus 4.5