project-idea-brainstorm
Pass
Audited by Gen Agent Trust Hub on Aug 12, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill fetches content from external sources such as Reddit and X (formerly Twitter) to inform the brainstorming process, which introduces a surface for indirect prompt injection.
- Ingestion points: Data enters the agent context via
WebSearchandWebFetchoperations targeting social media platforms in Step 2. - Boundary markers: The instructions do not specify any delimiters or safety markers (e.g., 'ignore embedded instructions') for handling the fetched external content before it is processed by the model.
- Capability inventory: The skill possesses capabilities to perform network searches (
WebSearch), fetch remote content (WebFetch), and write to the local file system (saving shortlists toproject-ideas-<date>.md). - Sanitization: There are no instructions for sanitizing, escaping, or validating the external content retrieved from social media threads before it is interpolated into the generation prompt in Step 3.
Audit Metadata