executing-plans
Pass
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to load and execute tasks from an external implementation plan file, which represents a surface where data could contain instructions that influence agent behavior.
- Ingestion points: Implementation plan files read in SKILL.md (Step 1).
- Boundary markers: Not explicitly defined, although the skill requires a 'review critically' phase before execution.
- Capability inventory: The skill performs file writes, git operations, and executes verification commands based on the plan's contents.
- Sanitization: The process requires identifying concerns and raising them with the user before implementation begins, along with a 'Hard Gate' checklist that mandates evidence for success before reporting completion.
Audit Metadata