devpilot-product-research
Pass
Audited by Gen Agent Trust Hub on May 2, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: Fetches market and competitor data from the internet via WebSearch tools to provide analysis on user pain points and trends.
- [COMMAND_EXECUTION]: Uses git commands (
git add,git commit) to stage and record research findings and rejected ideas to the local repository. - [PROMPT_INJECTION]: Ingests untrusted content from the web which is then used to synthesize product recommendations, creating a surface for indirect prompt injection.
- Ingestion points: WebSearch results processed by parallel sub-agents in SKILL.md.
- Boundary markers: Not explicitly implemented for the interpolated web data.
- Capability inventory: File write access to the
docs/directory and git execution for committing changes. - Sanitization: No explicit sanitization or filtering of external content before processing in the synthesis phase.
Audit Metadata