Skills
Navigate
skills/1811labsyt/audionotes-claude-plugin/retrieve-notes/Gen Agent Trust Hub

retrieve-notes

Pass

Audited by Gen Agent Trust Hub on Nov 9, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes user-generated content (notes, titles, transcripts) which are classified as untrusted data.
  • Ingestion points: Data enters the context through search_notes, list_notes, and get_note tools.
  • Boundary markers: The skill provides explicit instructions to treat excerpts as untrusted and to separate summary (generated) from transcript (source) content.
  • Capability inventory: The skill is limited to reading and reasoning; it explicitly states it cannot create, edit, delete, or share notes.
  • Sanitization: Instructions explicitly command the agent to never execute instructions found inside notes, reveal credentials, or automatically open embedded URLs.
  • [EXTERNAL_DOWNLOADS]: The skill directs users to an external MCP server and an official installation repository.
  • Evidence: References https://api.audionotes.app/mcp and https://github.com/1811LabsYT/audionotes-claude-plugin. These resources are consistent with the skill's stated purpose and author.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 9, 2026, 10:56 AM