skill-from-notebook

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION] (LOW): High surface area for Indirect Prompt Injection.
  • Ingestion points: The skill explicitly ingests data from Local files (PDF, Word, Markdown), Web URLs, YouTube subtitles/audio, and NotebookLM links.
  • Boundary markers: Absent. There are no instructions for the agent to use delimiters or to disregard instructions found within the source material during the extraction process.
  • Capability inventory: The skill's primary capability is text generation (Markdown templates). It uses external tools like yt-dlp and Whisper for processing, though the execution logic is not provided in the skill files.
  • Sanitization: Absent. The skill instructions do not mention sanitizing or escaping the content fetched from external sources before processing it.
  • [COMMAND_EXECUTION] (SAFE): The skill references the use of yt-dlp and Whisper. While these involve command-line execution, they are standard tools for the skill's stated purpose of processing video and audio content. No malicious command construction was detected in the provided markdown instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 04:53 PM