30x-growth-marketing-panel

Pass

Audited by Gen Agent Trust Hub on May 19, 2026

Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The documentation references an automation pipeline that utilizes CLI tools such as yt-dlp and notebooklm-py to process marketing data. These tools are described as being used for the skill's core purpose of knowledge distillation and expert persona creation.
  • [EXTERNAL_DOWNLOADS]: The skill is designed to retrieve data from YouTube to populate its expert knowledge base. This behavior is transparently documented and is central to the skill's primary function of providing expert advice based on existing content.
  • [PROMPT_INJECTION]: The skill processes external data (YouTube transcripts), which creates a surface for indirect prompt injection. This risk is managed through an 'Anti-Hallucination Protocol' and persona-based verification.
  • Ingestion points: YouTube transcripts and video metadata fetched via yt-dlp (referenced as an internal tool).
  • Boundary markers: The skill uses [Retrieved content from NotebookLM] markers and ⚠️ symbols to distinguish between retrieved facts and AI-generated extrapolations.
  • Capability inventory: Execution of CLI tools (yt-dlp, notebooklm-py) and API interactions with NotebookLM to manage the knowledge base.
  • Sanitization: The skill relies on retrieval-first logic and dual-verification with an 'Expert KB' to ensure data integrity.
Audit Metadata
Risk Level
SAFE
Analyzed
May 19, 2026, 09:18 AM
Security Audit — agent-trust-hub — 30x-growth-marketing-panel