aa15d95986
Adds lightweight AI governance extension with: - R6 workflow formalism (Rules, Role, Request, Reference, Resource → Result) - Session start/end hooks for session tracking and audit initialization - Before/after tool call hooks for R6 request logging and audit trails - /audit command for generating governance reports - Hash-linked audit chain for verifiable provenance The extension provides transparent accountability for AI agent actions without impeding productivity, using a policy-based approach that can be customized per team or organization. Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
24 lines
471 B
JSON
24 lines
471 B
JSON
{
|
|
"id": "web4-governance",
|
|
"channels": [],
|
|
"configSchema": {
|
|
"type": "object",
|
|
"additionalProperties": false,
|
|
"properties": {
|
|
"auditLevel": {
|
|
"type": "string",
|
|
"enum": ["minimal", "standard", "verbose"],
|
|
"default": "standard"
|
|
},
|
|
"showR6Status": {
|
|
"type": "boolean",
|
|
"default": true
|
|
},
|
|
"actionBudget": {
|
|
"type": ["number", "null"],
|
|
"default": null
|
|
}
|
|
}
|
|
}
|
|
}
|