subagent-orchestration

Pass

Audited by Gen Agent Trust Hub on Sep 6, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill provides behavioral guidelines for orchestrating subagents, advocating for fresh context and isolation per task, which helps prevent context-based prompt injection and data leakage.
  • [SAFE]: The recommended two-stage review cycle (specification compliance and code quality) serves as an effective security control to verify subagent outputs.
  • [SAFE]: Instructions for handoff protocols explicitly restrict the passage of implementation details or assumptions, ensuring subagents operate within a limited, well-defined scope.
  • [SAFE]: No evidence of credential exposure, unauthorized network activity, or obfuscation was found in the documentation or logic.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 6, 2026, 11:58 PM
Security Audit — agent-trust-hub — subagent-orchestration