muapi-media

Pass

Audited by Gen Agent Trust Hub on Aug 28, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill correctly manages sensitive credentials by accessing the MUAPI_API_KEY from the environment, avoiding hardcoded secrets or command-line exposure.
  • [SAFE]: The inclusion of a mandatory --confirm-paid flag in the generation process provides a security control against unintended financial costs incurred by autonomous agent actions.
  • [SAFE]: The download_output function implements a protocol check, ensuring that only HTTPS URLs are accessed when downloading generated media artifacts.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes an external model catalog to resolve request endpoints, creating a surface for indirect prompt injection. \n
  • Ingestion points: The model catalog is retrieved from the MUAPI_BASE_URL in scripts/muapi_media.py. \n
  • Boundary markers: No explicit delimitation or instruction-ignoring markers are used for the catalog data. \n
  • Capability inventory: The tool possesses network request capabilities (POST/GET) and file system write access. \n
  • Sanitization: The script validates that endpoints begin with /api/v1/ and that download links use HTTPS.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 28, 2026, 12:59 PM
Security Audit — agent-trust-hub — muapi-media