How do I install the Run One Spec skill?

Run npx skillmds@latest add liberatedpixelcup/run-one-spec in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Run One Spec safe to use?

This skill has not completed SkillMD's automated safety review yet. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Run One Spec?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Run One Spec free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published Run One Spec?

liberatedpixelcup (@liberatedpixelcup) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Run One Spec by liberatedpixelcup — Agent Skill

Name: Run One Spec
Author: liberatedpixelcup

Run One Spec

Run a single Node, browser, or Playwright spec instead of the full suite. Use when debugging one test, iterating on a failing spec, or the user asks to run one file, one test, or a filtered suite. Do not start with npm test.

File contents

Run one spec

Pick the suite the file belongs to. Do not start with npm test.

Isolation is below. Coverage is a separate full-suite run — see write-spec after a gated sources/ or scripts/ edit (skip if the path is in codecov.yml ignore:).

Node (`tests/node/`)

npm run test:node collects every *_spec.js / *_spec.ts and ignores extra args. Call node --test on the file instead:

node --test tests/node/scripts/generateSources/items_spec.ts
node --test --test-name-pattern="parseItem" tests/node/scripts/generateSources/items_spec.ts

npm run test:node:coverage also runs every Node spec, not this file alone.

Browser (Testem + Mocha)

tests/tests.js imports every browser spec. There is no CLI file filter. Do not comment out imports. A spec that is not imported never runs; Codecov patch then fails with no Mocha failure.

Filter with ?grep= (best isolation). Start the server, then open the harness with a grep pattern:

npm run test:server
# then open, in the launched browser:
# http://localhost:7357/tests_run.html?debug=false&grep=CategoryTree

Mocha's browser bundle reads grep, fgrep, and invert from location.search inside mocha.run(), which tests/tests.js calls. Every spec file is still imported, so the filter selects which tests execute, not which files load. The Mocha UI search box does the same thing by rewriting this query string.

If 7357 is busy, testem.cjs reads TESTEM_PORT:

TESTEM_PORT=7360 npm run test:server

it.only / describe.only works locally as a fallback. Do not commit it.

testem.cjs before_tests always runs every Node spec first, including this Chrome-only command. It is still the full Mocha browser suite, not one file:

node ./node_modules/testem/testem.js ci --launch Chrome

npm run test:browser:coverage is that full suite in Chrome and Firefox.

DEBUG=1 (or DEBUG=true) keeps the app's verbose debug output instead of the default ?debug=false.

Visual (Playwright)

npm run test:visual uses Playwright’s webServer (npm run dev locally). npx playwright test with this repo’s config does the same.

npx playwright test tests/visual/home.spec.js
npx playwright test -g "title substring"

Wait for catalog/canvas via tests/visual/home-helpers.ts. Do not register visual specs in tests/tests.js. They are not Codecov-instrumented. Argos upload needs ARGOS_TOKEN. When a change warrants a visual run: visual-test.

---name: run-one-specdescription: Run a single Node, browser, or Playwright spec instead of the full suite. Use when debugging one test, iterating on a failing spec, or the user asks to run one file, one test, or a filtered suite. Do not start with npm test.---# Run one specPick the suite the file belongs to. Do not start with `npm test`.Isolation is below. Coverage is a **separate** full-suite run — see[write-spec](../write-spec/SKILL.md#confirm-the-new-lines-are-hit) after agated `sources/` or `scripts/` edit (skip if the path is in[`codecov.yml`](../../../codecov.yml) `ignore:`).## Node (`tests/node/`)`npm run test:node` collects every `*_spec.js` / `*_spec.ts` and ignores extraargs. Call `node --test` on the file instead:```bashnode --test tests/node/scripts/generateSources/items_spec.tsnode --test --test-name-pattern="parseItem" tests/node/scripts/generateSources/items_spec.ts````npm run test:node:coverage` also runs every Node spec, not this file alone.## Browser (Testem + Mocha)[`tests/tests.js`](../../../tests/tests.js) imports every browser spec. There isno CLI file filter. Do not comment out imports. A spec that is not importednever runs; Codecov patch then fails with no Mocha failure.**Filter with `?grep=` (best isolation).** Start the server, then open theharness with a grep pattern:```bashnpm run test:server# then open, in the launched browser:# http://localhost:7357/tests_run.html?debug=false&grep=CategoryTree```Mocha's browser bundle reads `grep`, `fgrep`, and `invert` from`location.search` inside `mocha.run()`, which[`tests/tests.js`](../../../tests/tests.js) calls. Every spec file is stillimported, so the filter selects which tests execute, not which files load.The Mocha UI search box does the same thing by rewriting this query string.If 7357 is busy, `testem.cjs` reads `TESTEM_PORT`:```bashTESTEM_PORT=7360 npm run test:server````it.only` / `describe.only` works locally as a fallback. Do not commit it.`testem.cjs` `before_tests` always runs every Node spec first, including thisChrome-only command. It is still the full Mocha browser suite, not one file:```bashnode ./node_modules/testem/testem.js ci --launch Chrome````npm run test:browser:coverage` is that full suite in Chrome and Firefox.`DEBUG=1` (or `DEBUG=true`) keeps the app's verbose debug output instead of thedefault `?debug=false`.## Visual (Playwright)`npm run test:visual` uses Playwright’s `webServer` (`npm run dev` locally).`npx playwright test` with this repo’s config does the same.```bashnpx playwright test tests/visual/home.spec.jsnpx playwright test -g "title substring"```Wait for catalog/canvas via [`tests/visual/home-helpers.ts`](../../../tests/visual/home-helpers.ts).Do not register visual specs in `tests/tests.js`. They are notCodecov-instrumented. Argos upload needs `ARGOS_TOKEN`. When a changewarrants a visual run: [visual-test](../visual-test/SKILL.md).

Frequently asked questions