메뉴
Write a test to improve code coverage. Uses `cargo llvm-cov` to identify uncovered lines, then writes a focused test to cover them. Use when user says "write test", "improve coverage", "cover lines", or "write-test".
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
Reference for the ui-tests infrastructure: test suites, directory layout, test.toml format, snapshot workflow, and common commands. Use when working with ui-tests cases, debugging test failures, or scaffolding new tests.
Runs the cheadergen verification workflow in stages - lint, format, generation tests, and full test suite. Use when user says "verify", "run checks", "run tests", or wants to validate changes before committing.
Guidelines for writing Rust documentation. Refer to this when you need to write doc comments on functions, types or modules.
Look up Rust dependency API docs locally. Use when the user or task requires information about a dependency's API, types, traits, functions, or modules — e.g. "look up guppy's PackageGraph", "what methods does rustdoc_types::Item have", "show me clap's derive API".
| name | write-test |
| description | Write a test to improve code coverage. Uses `cargo llvm-cov` to identify uncovered lines, then writes a focused test to cover them. Use when user says "write test", "improve coverage", "cover lines", or "write-test". |
Write a single focused test to improve code coverage, guided by cargo llvm-cov output.
If the user specified a target file, run:
just uncovered <file>
If no file was specified, run just uncovered to see all uncovered lines, then pick the highest-impact target (most uncovered lines or most important code path).
IMPORTANT: After the initial scan, always scope just uncovered to the target file to keep context small.
Read the target file to understand what the uncovered lines do. Focus on understanding the code paths that lead to those lines — what inputs, configurations, or conditions trigger them.
Write a single focused test that exercises the uncovered code path. Place it in the appropriate test module or file following existing project conventions.
Only write one test per invocation. Keep it minimal and targeted at the specific uncovered lines.
Run coverage again and compare:
just uncovered <file>
Compare the before and after just uncovered output. Confirm that previously uncovered lines are now covered. If the test didn't improve coverage as expected, investigate why and adjust.
just uncovered output is compact (one line per file with collapsed ranges) to minimize token usage. Prefer it over reading raw lcov.info.