What problem does it solve? AI agents that call external tools can execute dangerous commands, leak sensitive data, or exceed intended boundaries without oversight. This Skill provides reusable patterns for enforcing policies, detecting malicious intent, tracking trust, and auditing every agent action. ## Core Features & Use Cases - Governance Policies: Define composable allowlists, blocklists, content filters, and rate limits as YAML configuration with most-restrictive-wins composition. - Intent Classification & Tool Decorators: Detect prompt injection, data exfiltration, and privilege escalation before execution, and wrap any tool function with a @govern decorator. - Trust Scoring & Audit Trails: Track agent reliability with decaying trust scores and maintain append-only audit logs exportable as JSONL. - Use Case: A support agent built with PydanticAI needs to search documents and create tickets but must never access shell commands or process credit card numbers — apply the governance decorator and content filters to enforce this automatically. ## Quick Start Add governance controls to my PydanticAI agent so it can only call approved tools, blocks sensitive patterns, and logs every action to an audit trail.
