extract-ideas
Pass
Audited by Gen Agent Trust Hub on Mar 6, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill processes untrusted external data which can contain malicious instructions intended to subvert the agent's logic (Indirect Prompt Injection).
- Ingestion points: The skill reads YouTube transcripts, newsletter content, and web trends provided by the orchestrator or fetched via web search (SKILL.md, Step 1; substack-notes-ideation.md, Step 1).
- Boundary markers: There are no explicit instructions to treat source material as data only or to ignore embedded instructions within the text.
- Capability inventory: The skill uses the Agent tool to run sub-tasks in the background and calls other skills such as creator-stack:voice and creator-stack:brand-guidelines (SKILL.md, Voice Application/Brand Compliance).
- Sanitization: No sanitization or validation logic is present to filter out potential prompt injection attacks hidden within the processed transcripts or web content.
Audit Metadata