shopping-shorts

Pass

Audited by Gen Agent Trust Hub on Oct 26, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [SAFE]: The skill uses cryptographic SHA-256 hash locking to bind approved concepts and storyboards to subsequent execution steps, preventing unauthorized modification of data during the production pipeline.
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted external data (product URLs and descriptions). It effectively mitigates this risk by instructing the agent to treat such content as data only and by enforcing multiple independent validation gates and human-in-the-loop review checkpoints.
  • [COMMAND_EXECUTION]: Local Node.js scripts execute the external binaries ffmpeg and ffprobe. These operations are secured through robust path resolution logic that strictly forbids directory traversal and identifies symbolic link path segments to ensure only authorized local media files are processed.
  • [SAFE]: The harness includes a credential scanning utility in 'scripts/prepare-render-plan.mjs' that automatically detects and blocks the inclusion of sensitive fields like API keys, bearer tokens, or secrets in provider request configurations.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 26, 2026, 10:43 AM