video-content-analyzer

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [Indirect Prompt Injection] (LOW): The skill is designed to ingest and process untrusted data from external social media platforms, creating a surface for indirect prompt injection.
  • Ingestion points: Processes caption, text, desc, and title fields from outliers.json (originating from Instagram, TikTok, and YouTube).
  • Boundary markers: The documentation does not specify the use of delimiters or 'ignore' instructions when passing this data to the Gemini AI.
  • Capability inventory: The associated script performs file read/write operations and makes network requests to the Gemini API.
  • Sanitization: There is no mention of sanitization or validation of the ingested video metadata before it is sent to the LLM.
  • [Credentials Unsafe] (SAFE): The skill correctly uses an environment variable (GEMINI_API_KEY) for authentication rather than hardcoding secrets.
  • [External Downloads] (SAFE): The skill relies on standard, well-known Python packages (google-genai, requests) and does not perform suspicious remote script execution.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 05:37 PM