Skills
Navigate
skills/starchild-ai-agent/official-skills/byok-custom-model/Gen Agent Trust Hub

byok-custom-model

Pass

Audited by Gen Agent Trust Hub on Sep 17, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied API documentation examples to assist in model registration.
  • Ingestion points: User-provided strings are ingested via the parse_example function in exports.py.
  • Boundary markers: The SKILL.md instructions require the agent to present a draft to the user for verification before any registration occurs.
  • Capability inventory: The skill is capable of updating the local custom_models.yaml configuration and triggering an internal cache refresh through the agent's management interface.
  • Sanitization: The skill utilizes integrated platform validation logic to verify and sanitize model parameters, URLs, and capabilities before persistence.
  • [SAFE]: The skill adheres to secure secret management standards by strictly enforcing the use of the request_env_input tool for API keys and bypassing chat for credential handling. Its internal management functions are isolated to loopback HTTP requests, ensuring secure local-only execution.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 17, 2026, 04:28 AM
Security Audit — agent-trust-hub — byok-custom-model