ns-investigator
Pass
Audited by Gen Agent Trust Hub on Sep 22, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill is consistent with its stated purpose, using standard commands for code investigation. It instructs the agent to proactively identify security risks like authentication bypasses and sensitive data exposure during its analysis.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data (logs, CI output, stack traces) which is a known attack surface for influencing agent behavior via malicious data. However, this is intrinsic to the skill's core functionality and is addressed by its evidence-based requirements.\n
- Ingestion points: SKILL.md (Workflow Step 1) ingests logs, traces, and test output from the user's environment.\n
- Boundary markers: No delimiters are defined for untrusted input.\n
- Capability inventory: The skill uses
git status,git diff, and search tools, along with executing local test and lint commands.\n - Sanitization: No specific sanitization of external content is mentioned.
Audit Metadata