ask-security-sentinel
Pass
Audited by Gen Agent Trust Hub on Feb 16, 2026
Risk Level: LOWPROMPT_INJECTION
Full Analysis
- Indirect Prompt Injection (LOW): The skill's function of scanning untrusted source code provides a surface for indirect prompt injection. Malicious instructions in scanned files could influence agent behavior.
- Ingestion points: Source code processed via 'verify code safety' triggers.
- Boundary markers: No isolation markers are defined.
- Capability inventory: Scripts are placeholders; actions are restricted to 'HALT' and 'warn'.
- Sanitization: No input validation or filtering is present.
Audit Metadata