skills/calesthio/openmontage/comfyui/Gen Agent Trust Hub

comfyui

Pass

Audited by Gen Agent Trust Hub on Aug 14, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill instructions require the agent to ingest and process community-sourced ComfyUI workflows, which are provided as JSON data.
  • Ingestion points: The agent accepts workflow_json or workflow_path as inputs (found in SKILL.md).
  • Boundary markers: The instructions lack specific guidance on using delimiters or safety warnings when interpolating external workflow data into prompts.
  • Capability inventory: The skill allows the agent to trigger image/video generation tools and write to the local file system (e.g., placing models in specific directories).
  • Sanitization: There is no mention of validating or sanitizing the content of community recipes or JSON workflows before processing.
  • [EXTERNAL_DOWNLOADS]: The skill explicitly guides the agent to facilitate the download of external assets, including models and software extensions (custom nodes).
  • Evidence: The skill instructs the agent to use "workflow author's model links" and download missing models via provided URLs (found in SKILL.md).
  • Evidence: It mentions installing "custom nodes" through the ComfyUI Manager or external documented paths, which involves executing code from third-party sources.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 14, 2026, 08:03 AM
Security Audit — agent-trust-hub — comfyui