the-crucible

Complexity analyst. Use when evaluating code complexity — identifies duplication, deep nesting, verbose patterns.

Comment analyst. Use when evaluating comments — identifies noise, redundancy, missing docs.

Documentation analyst. Use when evaluating API documentation — identifies missing, outdated, or insufficient docs.

Error handling analyst. Use when evaluating how errors are created, propagated, and handled. Does not fix, only reports findings.

Naming analyst. Use when evaluating identifier names — reports variables, functions, types that fail to reveal behavior and purpose.

Observability analyst. Use when evaluating logging, metrics, and tracing practices. Does not fix, only reports findings.

Security analyst. Use when evaluating code for common security issues — secrets, injection, input validation, crypto. Does not fix, only reports findings.

Test suite analyst. Use when evaluating test quality — checks if tests are lean, test the right things, and use proper strategies. Does not fix, only reports findings.

Multi-agent code quality evaluation — orchestrates specialized review agents across naming, complexity, tests, security, and documentation dimensions. Invoke whenever task involves any interaction with code quality — reviewing implementations, evaluating pull requests, or assessing code before commits.

Task completion validation — verify deliverables match original requirements. Invoke whenever task involves completing work for the user — code changes, documentation, research, refactoring, or any deliverable.

Standardized output format for code review agents. Provides consistent Critical/Issues/Recommendations structure for automated parsing and aggregation.