ideate-notes
Pass
Audited by Gen Agent Trust Hub on Mar 6, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: Fetches content from well-known services including YouTube and Substack archive pages to use as source material for ideation.
- [COMMAND_EXECUTION]: Utilizes YouTube MCP tools (search_videos, get_video_details, get_video_transcript, get_video_comments) to interact with external data sources.
- [PROMPT_INJECTION]: Vulnerable to indirect prompt injection due to the ingestion of untrusted third-party data.
- Ingestion points: The skill processes YouTube transcripts, top comments from videos, and Substack newsletter content in Steps 1 and 2.
- Boundary markers: There are no explicit instructions or delimiters defined to prevent the agent from following malicious instructions potentially embedded within transcripts or user comments.
- Capability inventory: The skill possesses the ability to read and write to the local filesystem (./substack/notes/), perform web searches, and execute MCP tools.
- Sanitization: The skill does not describe any sanitization, filtering, or validation of the fetched external text before it is passed to the ideation engine.
Audit Metadata