# Coverage

> Coverage is an Agent Skill for testing & quality published in the firedancer-io/firedancer repository on GitHub: a SKILL.md instruction file an AI agent can load to do this task.

Export source coverage of test

- Skill name: `coverage`
- Category: [Testing & Quality](https://www.openparable.com/skill/testing-quality/1)
- Source: https://github.com/firedancer-io/firedancer/blob/HEAD/.claude/skills/coverage/SKILL.md
- Repository: firedancer-io/firedancer (1,505 GitHub stars when collected)

## SKILL.md

Export source coverage of unit test $0:

1. Build: `make -j CC=clang EXTRAS=llvm-cov BUILDDIR=clang-cov $0`
2. Run: `CLANKER=1 ./contrib/test/single_test_cov.sh build/clang-cov/unit-test/$0`
   - produces `default.profdata`
3. Export
   - Per file: `llvm-cov show --instr-profile default.profdata build/clang-cov/unit-test/$0 -sources <src/foo/bar.c> -show-instantiations=false`
   - Per function: `llvm-cov show --instr-profile default.profdata build/clang-cov/unit-test/$0 --name=<function_name>`
   - All files overview (verbose!) `llvm-cov report --instr-profile default.profdata build/clang-cov/unit-test/$0`

If no further context/instructions are given, attempt to find under-covered areas of code:
Skim the test code to understand what it's trying to test, then drill down the cov report to see if it actually covers that code.

---

Source: https://www.openparable.com/skill/testing-quality/coverage (OpenParable). Cite the canonical URL above.
