artifact-resurfacing
Pass
Audited by Gen Agent Trust Hub on Nov 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from local documentation files (CLAUDE.md, MEMORY.md) to perform citation audits. This ingestion creates a surface for indirect prompt injection if those files contain untrusted content. The risk is minimized by the skill's use of read-only search commands, double-quoting of shell variables, and a strict 'propose-not-apply' governance protocol that requires manual verification of all proposed changes.- [SAFE]: The skill is a well-structured tool for repository maintenance. It lacks network access, privilege escalation, or obfuscation, and focuses on local file consistency checks with strong human-in-the-loop controls.
Audit Metadata