managed-model-endpoints

Pass

Audited by Gen Agent Trust Hub on Jul 17, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill provides legitimate guidelines for managing model endpoints and infrastructure without introducing malicious code or configurations.
  • [SAFE]: The skill promotes safe data handling by explicitly instructing the agent to store secret references as environment variables or credential references rather than in plain-text metadata, which follows standard security practices.
  • [PROMPT_INJECTION]: The workflow includes a step to capture model inference responses for verification, which creates a potential surface for indirect prompt injection. Ingestion points: Inference response data as described in SKILL.md workflow step 5. Boundary markers: No explicit delimiters for external content are specified. Capability inventory: Description mentions startup scripts and inference routing. Sanitization: No specific filtering or sanitization of the captured responses is mentioned. This risk is assessed as safe given the limited scope of the auditing task.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 17, 2026, 04:04 AM
Security Audit — agent-trust-hub — managed-model-endpoints