project-memory-resume
Pass
Audited by Gen Agent Trust Hub on Sep 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes project data from local markdown files (activeContext.md and session handoffs), which represents an indirect prompt injection surface.
- Ingestion points: Local file reads from the project-memory/ directory specified in SKILL.md.
- Boundary markers: The skill process requires a mandatory trust check and presenting a summary to the user for explicit confirmation before proceeding.
- Capability inventory: The skill is restricted to file reading and git status checks, with no write capabilities or dynamic execution.
- Sanitization: Instructions mandate treating external data as unconfirmed and requiring verification of existing blockers or checkboxes.
Audit Metadata