trim-cot-leakage
Pass
Audited by Gen Agent Trust Hub on Aug 20, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill provides specific ripgrep (rg) command templates in references/recall-batteries.md to search the local file system for prose patterns. These commands include flags like --hidden and --glob to control search scope within the repository environment.
- [PROMPT_INJECTION]: The skill involves reading and processing various files within a project scope (Markdown, code comments, etc.), which constitutes an ingestion point for potentially untrusted data.
- Ingestion points: Repository files processed during the audit workflow defined in SKILL.md.
- Boundary markers: Absent; the skill does not specify delimiters for the ingested text.
- Capability inventory: File system search (rg), file read, and file write access for auditing and remediation.
- Sanitization: Absent; the skill relies on semantic analysis of search hits to determine if they should be modified or deleted.
Audit Metadata