youtube-thumbnail

Pass

Audited by Gen Agent Trust Hub on Nov 6, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONDYNAMIC_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes external, untrusted data to generate thumbnail content, creating an attack surface where malicious input could influence command parameters or rendering behavior.\n
  • Ingestion points: Video titles and user-provided 'hooks' are read from the environment or input and used in processing (SKILL.md).\n
  • Boundary markers: Absent; the instructions do not specify delimiters or prompts to ignore instructions embedded in the metadata.\n
  • Capability inventory: The skill utilizes shell execution for ffmpeg, google-chrome (headless), and sips, and runs a local Python script (SKILL.md).\n
  • Sanitization: Absent; there are no instructions for sanitizing, escaping, or validating the input before its use in shell commands or the HTML template.\n- [COMMAND_EXECUTION]: The skill instructs the agent to execute multiple command-line tools for media processing and validation.\n
  • Evidence: Subprocess execution of ffmpeg for frame extraction, google-chrome for rendering, sips for format conversion, and a local script youtube_check.py.\n- [DYNAMIC_EXECUTION]: The skill performs runtime generation of visual assets by dynamically populating an HTML template (assets/thumbnail-template.html) and rendering it through a browser engine to produce a final image.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 6, 2026, 07:39 PM