The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes standard development commands to facilitate the code review process.
Evidence: It uses git diff main...HEAD to retrieve changes and npm scripts (such as npm run test:run, npm run test:coverage, and npm run typecheck) to verify code compliance and test adequacy.
[SAFE]: The skill incorporates comprehensive security and quality guidelines that follow industry best practices.
Evidence: The file references/security-checklist.md implements a detailed review framework based on the OWASP Top 10 patterns, including checks for injection, credential exposure, and access control failures.
Evidence: The skill defines an adversarial posture and provides a structured rationalization-refutation.md reference to prevent the agent from accepting common excuses to skip quality or security gates.
[SAFE]: The skill uses an MCP-based orchestration system for sensitive operations, ensuring that state transitions and automated scanning are handled via the vendor's specialized tools.
Evidence: Use of exarchos_orchestrate and exarchos_workflow actions for security scanning (check_security_scan), static analysis, and workflow state management.

review