northstar-cleanup
Pass
Audited by Gen Agent Trust Hub on Sep 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied text (docs tree or cleanup questions) which serves as an ingestion point for untrusted data. Ingestion points: SKILL.md (text after /northstar-cleanup). Boundary markers: Present; the instructions explicitly mandate a read-only posture and require operator authorization for repairs. Capability inventory: Documentation inspection, triage, and potential modification of files. Sanitization: Not specified.
- [SAFE]: The skill follows security best practices by implementing default read-only constraints and prohibiting autonomous execution of high-risk tasks such as file deletion, worker initialization, or unauthorized merging.
Audit Metadata