brainstorm
Pass
Audited by Gen Agent Trust Hub on Jul 10, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: Indirect prompt injection surface identified through untrusted file ingestion. The skill reads existing game concept and pillar documents to allow users to resume previous brainstorming work, which could contain malicious instructions.
- Ingestion points: Files 'design/gdd/game-concept.md', 'design/gdd/game-pillars.md', and 'production/review-mode.txt'.
- Boundary markers: The skill does not employ boundary markers or specific instructions to ignore embedded commands within the read data.
- Capability inventory: The skill has 'Write' access to the filesystem to update design documents and can use the 'Task' tool to invoke other director agents with the ingested data.
- Sanitization: No sanitization, escaping, or validation of the contents of the read files is performed before processing.
Audit Metadata