architecture-review
Pass
Audited by Gen Agent Trust Hub on Sep 26, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill uses established Ontoly CLI commands for its intended purpose of repository analysis. No evidence of prompt injection, data exfiltration, or obfuscation was found.
- [INDIRECT_PROMPT_INJECTION]: The skill possesses an ingestion surface via repository file inspection fallback (documented in reference/fallbacks.md and reference/workflow.md). Boundary markers in SKILL.md and reference/fallbacks.md restrict this to cases where graph analysis is insufficient. Capabilities include the ontoly CLI tool suite. No specific sanitization is defined for the fallback content.
Audit Metadata