skill-coverage-audit

To map codepaths to tests, the coverage audit identifies changed files in your pull request, extracts control-flow and error path branches, and maps them against existing tests to highlight untested paths. It outputs a coverage diagram showing exactly which paths lack tests.

Yes, you can automatically generate missing tests for uncovered branches. The audit identifies untested codepaths and auto-generates targeted tests following your project conventions, producing up to 20 tests to fill coverage gaps before PR submission.

Tracing control-flow branches and error paths in staged changes is best handled by auditing the diff to extract a codepath inventory. It explores up to 30 distinct codepaths across changed files, analyzing control-flow and error branches within a 2-minute limit per path.

Yes, the test generation works with existing project conventions. It analyzes your codebase patterns and generates targeted tests for uncovered paths that follow the conventions already established in your project's test suite.

Limitations when running a coverage audit on a large pull request include enforced caps of 30 codepaths, 20 generated tests, and a 2-minute exploration limit per path. Very large diffs with extensive branching may exceed these thresholds and require manual review.

Untested codepaths remain because existing tests may not execute all control-flow branches or error paths within changed files. A coverage audit maps your diff to the test suite, identifying exactly which modified branches lack execution and auto-generating tests for those gaps.