capy-video-gen-skill

Warn

Audited by Gen Agent Trust Hub on Mar 21, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill dynamically loads Python classes at runtime using importlib.import_module based on strings provided in configuration files (tools/render_backend.py). It also utilizes libraries like moviepy and scenedetect which execute system commands (ffmpeg) for video processing and scene detection.
  • [EXTERNAL_DOWNLOADS]: The skill performs network requests to non-whitelisted domains to facilitate AI generation and content retrieval, including yunwu.ai, trickle-lab.tech, and openrouter.ai.
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it ingests untrusted text data and interpolates it into prompts for generation tools. 1. Ingestion points: novel_text in agents/event_extractor.py and script in agents/storyboard_artist.py. 2. Boundary markers: Absent; user-provided text is directly embedded in prompt templates. 3. Capability inventory: Network API calls for image/video generation and local file reading for asset processing. 4. Sanitization: No explicit filtering or validation of input text before construction of tool prompts.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Mar 21, 2026, 07:18 AM