Imran Siddique
|
dcfae78fa4
|
feat: add agent-governance skill
Add governance patterns and techniques for AI agent systems:
- Policy definition with allowlists, blocklists, and content filters
- Semantic intent classification for threat detection
- Tool-level governance decorator pattern
- Trust scoring with temporal decay for multi-agent systems
- Append-only audit trail design
- Framework integration examples (PydanticAI, CrewAI, OpenAI Agents)
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
|
2026-02-18 13:28:25 -08:00 |
|