prime
Pass
Audited by Gen Agent Trust Hub on Feb 16, 2026
Risk Level: LOW
Full Analysis
- Prompt Injection (SAFE): No instructions to override system behavior or safety filters were detected.
- Data Exposure & Exfiltration (SAFE): The skill does not access sensitive system files or credentials, and no network exfiltration paths are present.
- Indirect Prompt Injection (LOW): The skill ingests untrusted data from README.md. While this presents an attack surface where a malicious README could influence the agent's summary, the lack of side-effect tools (write/network) limits the risk to an informational level.
- Command Execution (SAFE): Bash execution is restricted to the git command suite, and the specific command used (git ls-files) is non-destructive.
Audit Metadata