new-skill
Pass
Audited by Gen Agent Trust Hub on Oct 16, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute local repository scripts such as npm run validate and npm run evals to verify the structure and schema of newly generated files.
- [COMMAND_EXECUTION]: It utilizes the cp command to deploy generated skills to the agent's local testing directory (~/.claude/skills/brooks-lint/) for verification in a live session.
- [INDIRECT_PROMPT_INJECTION]: The skill possesses a vulnerability surface for indirect prompt injection as it generates new instructions based on untrusted user input. 1. Ingestion points: User input for the skill name and purpose provided via $ARGUMENTS in SKILL.md. 2. Boundary markers: Absent in the generated SKILL.md and guide templates. 3. Capability inventory: Creation of new files in the skills directory, modification of evals.json, and shell command execution via npm. 4. Sanitization: No explicit sanitization or escaping of the user-provided inputs is required before they are interpolated into the generated files.
Audit Metadata