ai-builder-mentor
Pass
Audited by Gen Agent Trust Hub on Aug 31, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructions include a step to execute a local validation script (scripts/validate-mentor-review.py) to check the format of generated reviews. This is a controlled execution of a script provided within the skill package and does not involve untrusted remote code.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied plans for AI projects, which serves as an ingestion point for untrusted data.\n
- Ingestion points: User requests describing AI product ideas, workflows, and distribution strategies are processed in SKILL.md (Steps 1 and 2).\n
- Boundary markers: The agent is directed to use a structured markdown template (assets/mentor-review-template.md) to isolate its analysis.\n
- Capability inventory: The skill has the capability to execute a local validation script via a shell command (SKILL.md, Step 7).\n
- Sanitization: Step 1.2 includes instructions to reject requests for persona imitation, and Step 7 utilizes a validation script to ensure the output adheres to expected structure and safety guidelines.
Audit Metadata