rpg-character

Pass

Audited by Gen Agent Trust Hub on Sep 21, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill downloads content from fal.ai CDN URLs during the character generation pipeline. This is a core part of the functionality for retrieving generated concept images and 3D meshes.
  • [INDIRECT_PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection as it processes user-provided ideas and external image URLs.
  • Ingestion points: tools/pipeline/concept.py accepts a text idea via the --idea argument, and tools/pipeline/run_character.py accepts a URL via the --image-url argument.
  • Boundary markers: No explicit delimiters or instruction-ignore warnings are specified for the input text or URL.
  • Capability inventory: The pipeline involves executing local Python scripts, writing files to the local assets/ directory, and performing network requests to external AI services (fal.ai).
  • Sanitization: The instructions do not specify sanitization or validation logic for the input strings or remote URLs, though the URLs are typically sourced from the previous pipeline stage.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 21, 2026, 05:18 AM
Security Audit — agent-trust-hub — rpg-character