my-harness-workflow

Pass

Audited by Gen Agent Trust Hub on May 16, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill serves as a documentation-only resource providing domain-specific knowledge. It lacks executable logic, scripts, or automated triggers that perform sensitive operations.
  • [DATA_EXFILTRATION]: No patterns for unauthorized data access or network communication were identified. The file paths mentioned, such as .moai/specs/, are project-specific markers for requirement documents and do not involve system-level sensitive files.
  • [EXTERNAL_DOWNLOADS]: The skill does not reference any external URLs for downloading code, scripts, or configuration files.
  • [COMMAND_EXECUTION]: Shell commands appearing in the documentation, such as ls .claude/hooks/moai/handle-*.sh | wc -l, are provided as illustrative examples for verification steps within the workflow rather than instructions for autonomous execution by the agent.
Audit Metadata
Risk Level
SAFE
Analyzed
May 16, 2026, 10:57 AM
Security Audit — agent-trust-hub — my-harness-workflow