executing-plans
Pass
Audited by Gen Agent Trust Hub on Oct 1, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill provides a standard procedural workflow for technical implementation. No malicious code, unauthorized command execution, or data exfiltration patterns were detected.
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and execute instructions from a plan file, creating a surface for indirect prompt injection. 1. Ingestion points: External plan files (SKILL.md, Step 1). 2. Boundary markers: The skill mandates a critical review of the plan and human partner confirmation if concerns are identified (Step 1.3). 3. Capability inventory: The skill directs the agent to follow all plan steps exactly, which may include using other tools, scripts, or sub-skills. 4. Sanitization: None. While this surface exists, the presence of human-in-the-loop review checkpoints and the instructional requirement to 'stop and ask for help' align with the skill's primary purpose and mitigate the risk level.
Audit Metadata