ai-agent-design
Pass
Audited by Gen Agent Trust Hub on May 16, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is a static documentation file providing architectural guidance and code snippets for educational purposes.
- [PROMPT_INJECTION]: While the text contains phrases often associated with prompt injection (e.g., "ignore all previous instructions"), these occur exclusively within code examples demonstrating how to implement security guardrails or within sections explaining common security mistakes. They do not represent an attempt to override the agent's behavior.
- [EXTERNAL_DOWNLOADS]: No external downloads or remote script executions are present.
- [REMOTE_CODE_EXECUTION]: No remote code execution patterns were detected.
- [DATA_EXFILTRATION]: No sensitive data access or exfiltration patterns were detected.
Audit Metadata