agents-sdk

Pass

Audited by Gen Agent Trust Hub on Sep 11, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [Credential Management]: The skill references a specific platform tool (openai-platform-api-key) for handling API credentials. This is a secure practice as it avoids manual handling or hardcoding of sensitive tokens within the skill itself.\n- [Indirect Prompt Injection Surface]: The skill is designed to process user ideas and repository content to generate implementations. This creates a surface for indirect prompt injection if malicious instructions are embedded in the repository. The skill encourages defining strict agent contracts to help manage this surface.\n
  • Ingestion points: User input and repository files (SKILL.md).\n
  • Boundary markers: None identified.\n
  • Capability inventory: Instructs the agent to provide local run commands and implementation logic.\n
  • Sanitization: None identified.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 11, 2026, 11:52 PM
Security Audit — agent-trust-hub — agents-sdk