agency-multi-agent-systems-architect
Pass
Audited by Gen Agent Trust Hub on Jul 6, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious patterns or vulnerabilities were detected. The skill is instructional in nature and provides a comprehensive framework for building resilient and secure AI agent pipelines.
- [SAFE]: The instructions explicitly guide the agent to treat external content as hostile and implement rigorous defenses, including output schema validation and the use of sanitizer agents to extract structured data from untrusted sources.
- [SAFE]: The skill mandates the principle of least privilege, stating that agents should only have access to the specific tools and data required for their roles, and recommends sandboxing for any code execution capabilities.
- [SAFE]: No suspicious network operations, hardcoded credentials, obfuscation, or persistence mechanisms were found in the provided files.
Audit Metadata