self-improve
Pass
Audited by Gen Agent Trust Hub on Apr 12, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [SAFE]: The skill focuses on project-specific maintenance by auditing local SKILL.md and CLAUDE.md files. It utilizes standard tools like Read, Grep, and Glob and does not attempt network exfiltration or sensitive file access.
- [PROMPT_INJECTION]: The skill ingests instructions from other files, creating a surface for indirect prompt injection. Ingestion points: SKILL.md library and CLAUDE.md files. Boundary markers: none. Capability inventory: Read, Grep, Glob. Sanitization: none. This is low risk given the tool's restricted scope and intended maintenance purpose.
Audit Metadata