Supercov

Supercov runs the project's own test command on the user's machine and reports line, branch and MC/DC coverage per test. It changes no source, tests or configuration. This skill uses that report to pick one untested piece of code at a time and write a focused test for it. It works with JavaScript, TypeScript, Python, Ruby, Rust, Go, Java and Kotlin projects.

The workflow has four steps. First, measure the full suite under Supercov; coverage is measured locally and sends nothing over the network. Second, list the untested code, pick one item and read it in context. Third, write one focused test for that target, changing only test files unless the user asks otherwise. Fourth, rerun the command from step one and report the test you added with the coverage before and after, using the numbers Supercov printed.

Use it when the user asks to add or improve tests, asks which code is untested or wants to raise coverage, or asks which file to refactor first and has agreed to a quality check.

Do not change application code to raise coverage unless the user asks. Do not run the quality commands without the user's request and consent. For behavior that does not exist yet and where the test should come first, the related skill test-driven-development is the better fit. Remember what coverage means: it shows which code ran under a test, not that the test checked the result. Supercov measures only the suites listed in its supported-suites guide.

Preconditions are Node.js 22 or newer (tell the user if it is missing and do not install software without approval) and the project's test command, such as npm test, pytest or cargo test; ask if it is not obvious from the repository. Always run the pinned version, 3.0.4. The first run downloads that release from npm, so get the user's approval first. The test command runs as the project defines it, so any services, credentials or network access the suite needs still apply.

The outputs are one new test per gap, run data written to a .supercov directory in the project, and a before-and-after coverage report. Core commands:

npx supercov@3.0.4 -- npm test
npx supercov@3.0.4 runs latest gaps --limit 5
npx supercov@3.0.4 runs latest file src/session.js

Other languages use the same pattern with pytest, cargo test, go test ./... or ./gradlew test. For other topics, the docs command prints the guide for that version, with topics agent-loop, quality, supported-suites and troubleshooting.

Supercov installs from the npm registry as a pinned release, and upgrading the pin needs a new review. It wraps whatever test runner the project uses. The optional quality check ranks files by code smells and reports what a change introduced, but it sends the project's source files to TypeSafe's API. Run it only when the user asked for a quality check, and tell them first that their source files will be sent. It needs the user's own TYPESAFE_API_KEY in the environment; never ask for a key in the conversation and never create, read or print one. If the key is missing, say the check needs it, answer from reading the code, and state that Supercov did not check it. Coverage commands make no network requests.

Developers who want steady, reviewable coverage gains one test at a time, with a clear audit trail of numbers before and after. It suits teams that care about supply-chain caution, such as pinned versions, approval before downloads and a hard line against touching application code.