review-skill-improver
Pass
Audited by Gen Agent Trust Hub on Sep 20, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests external feedback logs which could contain malicious instructions designed to manipulate the 'Proposed Fix' section of the generated improvement report.
- Ingestion points: The skill reads feedback logs from repository paths or attached artifacts as specified in the 'Input' section of SKILL.md.
- Boundary markers: The skill implements 'Hard gates' for schema validation and input location, which manage structural integrity but do not address semantic content safety.
- Capability inventory: The skill is capable of reading file contents and generating Markdown reports containing suggested code changes for other skills in the repository.
- Sanitization: The 'Schema / shape' gate identifies and lists malformed rows for skipping, but there is no mechanism to filter or sanitize the semantic content of the rationales before they are included in recommendations.
Audit Metadata