What problem does it solve? Teams often lack a trustworthy, up-to-date picture of how much of their codebase tests actually cover, and coverage numbers get guessed, recalled from stale docs, or padded by low-quality tests. This Skill runs the project's own instrumented test suites once per stack, judges the real numbers against the user's coverage requirement, and records the verdict in a durable, reconcilable document. ## Core Features & Use Cases - Stack-aware tooling detection: Detects the coverage tooling each surface already uses (coverlet for .NET, the Angular builder's coverage flag, the declared runner for plain JS/TS) and never installs or picks a runner itself. - Measured, judged, documented: Runs the instrumented suite once per surface, keeps raw machine-readable output (cobertura.xml, lcov.info, coverage-summary.json), and writes a COVERAGE.md with per-stack verdicts, per-module tables, and weak points tiered as small, substantial, or structural. - Use Case: Ask how covered a mixed .NET API plus Angular frontend workspace is; the Skill measures each surface separately, compares against your 90% bar, and hands a ranked weak-point list to the coverage-fix loop. ## Quick Start Measure the test coverage of this project against a 90% line coverage bar and write the coverage report.
