Запуск тестов с кратким дайджестом результатов

★ 7.0 · testing

Run and Manage Tests is a Claude Code skill that executes clice test suites — unit, integration, smoke, snap, or all at once — with a configurable build type, and returns only a concise digest to the main conversation. Reach for it when you need to run the full test suite before pushing a commit, reproduce a specific failing case by filter, or quickly see which tests broke and exactly where — without flooding the chat with raw output. The skill runs in a forked context: all test output stays isolated, and the main dialog receives only a per-suite summary — pass/fail counts, the name of each failing test, file:line, the assertion message or diff excerpt, and the exact filter command to reproduce the failure. Execution uses pixi commands (pixi run unit-test, integration-test, smoke-test, snap-test, test) with Debug or RelWithDebInfo build types (default: RelWithDebInfo). Integration and snap suites rely on Vitest and npm; smoke tests use a Node replay.ts script. Suites always run sequentially — parallel execution is blocked because they share the tests/data workspace. This skill only runs and reports — it never modifies code or tests, skips checks, or sets UPDATE_SNAPSHOTS; any fixes happen in the main conversation with full context. Works within a pixi environment set up for the clice project; not applicable outside it.