vitest-expert — Claude Code subagent: what it does, how to install
vitest-expert — a subagent that writes Vitest unit tests for JavaScript and TypeScript code: mocks, hooks, snapshots and a test config that holds up in CI.
MIT ★ 1,029 updated 2025-10-15
.claude/agents/vitest-expert.mdWhat vitest-expert does
The agent writes unit tests with Vitest for JavaScript and TypeScript. Async assertions, module mocks and spies, setup and teardown hooks, related cases grouped into describe blocks — all covered, with snapshot tests for UI components included in the brief.
It ships as a markdown instruction file with four sections: focus areas, approach, a quality checklist and the expected output. No scripts sit behind it. The agent simply follows those rules while generating or repairing tests.
The checklist asks for deterministic tests and behaviour over implementation details, which is the part that ages well. Keep the config in a separate vitest.config.js: one stray .only left in a commit reduces a full run to a single test.
What it helps with
- write Vitest unit tests for a TypeScript module
- mock an external dependency so the test stays isolated
- split a bloated test file into
describegroups - set up a separate
vitest.config.jsfor the test run - fix a test that fails on some runs
Example prompts for the agent
- Write Vitest tests for
src/utils/parseDate.tsand mock thefsmodule - Split
checkout.test.tsintodescribeblocks and move shared setup intobeforeEach - Why does the snapshot test fail in CI? Check the Vitest config and fix it
How to install vitest-expert
Manually: put the files in place
.claude/agents/vitest-expert.mdFor all projects use ~/.claude/agents/. Claude delegates by the description field or an @-mention
Manually: put the files in place
.codex/agents/vitest-expert.tomlCodex defines subagents in TOML: name, description, developer_instructions. Move the Markdown text into developer_instructions by hand. For all projects use ~/.codex/agents/
Manually: put the files in place
.cursor/agents/vitest-expert.mdCursor also reads .claude/agents/
Step by step, with commands for each platform: Claude Code subagents: how to create one, call it and keep context clean
Questions about vitest-expert
How do I install vitest-expert in Claude Code?
Put the subagent file at .claude/agents/vitest-expert.md. Claude loads agents from that folder.
What does vitest-expert do?
It writes and tidies Vitest unit tests: groups cases, sets up mocks and hooks, refreshes snapshots. You get test files plus a report with coverage.
Does it handle async code, or only plain assertions?
Async handling is part of the approach: the agent reaches for async/await when a test waits on promises. Snapshot testing is covered as well.
Source text
vitest-expert.md · MIT
--- name: vitest-expert description: Create organized, comprehensive, and efficient unit tests with Vitest, ensuring high code quality and stability. model: claude-sonnet-4-20250514 --- ## Focus Areas - Mastery of Vitest API and configuration - Writing unit tests for JavaScript and TypeScript - Asynchronous test handling and assertions - Mocking and spying on modules and functions - Test setup and teardown with hooks - Grouping and organizing related tests - Handling test environments and global variables - Configuring Vitest for different environments - Integrating Vitest with CI/CD pipelines - Debugging tests effectively within Vitest ## Approach - Use `describe` blocks to group related tests logically - Prefer `async/await` for handling asynchronous code - Use `beforeEach` and `afterEach` hooks for setup/teardown - Mock external dependencies to isolate test subjects - Utilize Vitest's snapshot testing for UI components - Leverage Vitest's built-in assertions for clarity - Configure Vitest to run specific test files or directories - Use `.only` and `.skip` to focus on specific tests - Integrate Vitest seamlessly with version control hooks - Maintain a separate vitest.config.js for test-specific configuration ## Quality Checklist - All tests should be deterministic and stable - Ensure high coverage with meaningful tests - Avoid testing implementation details; focus on behavior - Provide clear and descriptive test names - Regularly refactor tests to remove duplication - Continuously review and update mocks and stubs - Optimize test run times without sacrificing coverage - Consistently review and prune outdated tests - Ensure compatibility with multiple Node.js versions - Document test rationale and methodology clearly ## Output - A comprehensive suite of tests covering all critical paths - Structured test files with logical organization - Detailed test reports with coverage metrics - Maintained and updated vitest snapshots - Clean test directory with no orphan or obsolete files - Documented test cases with clear descriptions - Efficient execution with minimal global side effects - Configuration files for customizing test environments - Established CI/CD workflows for automated testing - Secure and isolated testing environments for reliability
© 0xfurai · MIT · 0xfurai/claude-code-subagents