render-narrated-ugc-wardrobe-stitch

Pass

Audited by Gen Agent Trust Hub on Aug 10, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [UNVERIFIABLE_DEPENDENCIES_AND_REMOTE_CODE_EXECUTION]: The skill orchestrates video production by calling local shell scripts (.sh) and a Python script (build_edl.py). These components are used for deterministic video assembly and are expected parts of the local project environment.
  • [DATA_EXPOSURE_AND_EXFILTRATION]: The skill uses Playwright to capture a screenshot of a specific brand landing page (https://www.bioma.health/) and interacts with established AI platforms (ElevenLabs, FAL) for media synthesis. These network operations are aligned with the skill's stated purpose.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes Whisper-generated word timings and project configurations to create video captions and edit lists.
  • Ingestion points: Audio transcription files (vo-final.words.json) and configuration files referenced in SKILL.md.
  • Boundary markers: None identified.
  • Capability inventory: Execution of local assembly scripts, FFmpeg, and browser automation via Playwright.
  • Sanitization: Not applicable beyond the specified token re-spelling logic.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 10, 2026, 02:59 PM
Security Audit — agent-trust-hub — render-narrated-ugc-wardrobe-stitch