coverage-report

coverage-report

Numbers come only from the tool. The model never estimates, infers, or rounds coverage. The model's only role is joining tool output to the manifest's reasons and narrating the observations section.

Read ${CLAUDE_PLUGIN_ROOT}/rules/coverage-report.base.md and the repo's .claude/coverage/refs/coverage-manifest.yml first.

One command, any time

The whole report is one wrapper — no arguments needed (it auto-detects the .sln):

./.claude/coverage/tools/report.sh

It collects coverage, joins it against the manifest, and writes a dated per-run folder .claude/coverage/reports/<YYYY-MM-DD>/ containing REPORT.md, REPORT.html, and CANNOT-TEST.md, then prints the Markdown. It exits non-zero only for a check named in the manifest's gate.enforce (empty by default, so a breach prints as ADVISORY and the script still succeeds); a failing build or test run is fatal regardless, from run-coverage.sh. Each run writes its own dated folder rather than overwriting a fixed path, so prior dates are preserved for comparison; override the date with REPORT_DATE=YYYY-MM-DD to re-emit under a specific day. CANNOT-TEST.md is generated from the manifest cannot_test entries (grouped by blocking construct, each row citing target, reason, and unlock), not hand-written. This is the identical join CI runs — the report a developer sees locally is the report on the PR. When invoked as the coverage-report skill, just run it and surface the result; don't re-derive numbers by hand.

Steps

1. Run the wrapper. ./.claude/coverage/tools/report.sh (or, before init has copied it, ${CLAUDE_PLUGIN_ROOT}/scripts/report.sh). Pass a solution path or repo-filter only if auto-detect is wrong. If the tools or solution are missing, report that — never fabricate numbers. All numbers come from the tool + the deterministic join in coverage-gate.py; the model does not parse XML or compute percentages.

   The script prints KIT_VERSION=, KIT_DRIFT= and FILE_FILTER= before collecting. Check all three. KIT_DRIFT= is <state> <manifest kit_version> <kit semver>: any state other than current means this repo has not been reconciled against the current kit, and the report carries a note saying so. Surface it in one line and name coverage-redo as what applies it. Do not apply any migration here: a report that quietly changed the measured scope would make its own numbers unreproducible. The one exception is the stale-tool-copy refresh below, which changes no classification. See ${CLAUDE_PLUGIN_ROOT}/MIGRATIONS.md ("Preflight"). Stale tool copies no longer need asking about: report.sh runs kit-sync.py first, which installs the current copies and applies the auto manifest entries, printing [kit-sync] lines for each. Relay those lines and remind the user to COMMIT the changed files, since CI runs the committed copies and not their kit checkout. If the sync printed nothing, the repo was already current. If FILE_FILTER lacks an exclusion for a shared directory that is present in the repo, stop and fix scope.vendored_paths before trusting the number.