agent-governance
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill references the official Microsoft 'agent-governance-toolkit' repository as an external resource.
- [PROMPT_INJECTION]: Includes regular expression patterns designed to identify and block prompt injection and instruction override attempts.
- [COMMAND_EXECUTION]: Provides Python code examples demonstrating how to wrap tool functions with governance logic and implement path traversal checks.
- [DATA_EXFILTRATION]: Contains defensive regex patterns to detect data exfiltration attempts such as the use of 'curl -d' or 'export to external'.
Audit Metadata