reconcile
Pass
Audited by Gen Agent Trust Hub on Sep 3, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from the repository being analyzed. 1. Ingestion points: Git logs, diffs, and project files read during state and SSOT checks in SKILL.md. 2. Boundary markers: None present in the instructions. 3. Capability inventory: Git diagnostic commands and file system read access. 4. Sanitization: None.
- [SAFE]: No malicious patterns detected. The skill implements standard Git-based diagnostic workflows. External links reference the author's own repository and follow established safety patterns.
Audit Metadata