ecc-pr-test-analyzer

Map the changed functions, classes, and modules to their corresponding tests, then identify untested code paths. Focus on behavioral coverage: check that edge cases, error paths, and important integrations are defended by tests with meaningful assertions.

A good review reports only actionable findings tied to a file, line, triggering condition, and observable impact. It distinguishes verified defects from hypotheses, checks callers and existing tests before reporting gaps, and rates gaps by impact as critical, important, or optional.

No, the review is static and read-only. It does not edit files, run project code or tests, or connect to databases; it relies on read/search access and asks the caller for a diff or runtime evidence when tools cannot obtain it.

Static review cannot confirm runtime behavior, so findings about execution-dependent issues remain unverified hypotheses. When a required area cannot be assessed from the available code, the review explicitly states that rather than claiming approval.

Gaps are rated by impact: critical gaps block the change, important gaps should be addressed, and optional gaps are omitted from blocking findings. A test is recommended only when a plausible bug would actually fail it.