brainstorming
Pass
Audited by Gen Agent Trust Hub on Sep 30, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill exhibits an indirect prompt injection attack surface by ingesting untrusted data from the local environment to inform its brainstorming process. \n- Ingestion points: In SKILL.md, the workflow requires the agent to 'Explore project context — check files, docs, recent commits'. \n- Boundary markers: There are no instructions or delimiters provided to help the agent distinguish between project data and its own system instructions, nor is there an 'ignore embedded instructions' warning. \n- Capability inventory: The skill allows the agent to write new design documents to the docs/plans/ directory, commit them to the git repository, and transition to implementation by invoking the writing-plans skill. \n- Sanitization: The instructions do not define any sanitization, filtering, or validation steps for the data retrieved from the project context.
Audit Metadata