chromatic-visual-regression-testing
Authors and runs Chromatic visual tests on Storybook, Playwright, or Cypress projects via the `chromatic` CLI; configures baselines, TurboSnap, UI Review, and CI gating; reads exit codes for change-vs-error classification. Use when the project ships visual regression coverage to Chromatic Cloud.
Install with skills.sh (any agent)
npx skills add testland/qa --skill chromatic-visual-regression-testingchromatic-visual-regression-testing
Overview
Chromatic is a visual testing platform built by the team behind Storybook (chromatic-overview (opens in new window)). It works in three modes:
The CLI shape: chromatic is one binary; everything (auth, framework choice, TurboSnap) is a flag or env var.
When to use
Authoring
Install
npm install --save-dev chromatic(Per the Chromatic quickstart (opens in new window).)
Connect to a project
The first run uploads the Storybook build, captures a snapshot of each story, and establishes those snapshots as the project baseline (quickstart (opens in new window)):
npx chromatic --project-token <your-project-token>Subsequent runs compare new snapshots to the baseline; visual changes must be explicitly accepted in the Chromatic UI before the build passes.
Authoring stories (Storybook mode)
Chromatic snapshots whatever your Storybook indexes, so authoring is "add a story." Per-story knobs that affect snapshots are passed in the story's parameters.chromatic block (e.g. viewports, delay, pauseAnimationAtEnd).
Authoring tests (Playwright / Cypress mode)
When invoked with --playwright or --cypress, the CLI captures snapshots from your existing Playwright / Cypress test runs (cli (opens in new window)); no per-test code changes are required beyond running the matching framework's archive output through Chromatic.
Running
CLI invocation
The canonical invocation in CI, with the token sourced from an env var (cli (opens in new window)):
export CHROMATIC_PROJECT_TOKEN=<token>
npx chromaticUseful flags (cli (opens in new window)):
| Flag | Effect |
|---|---|
--project-token <t> | Override the token (env var preferred). |
--playwright | Capture from Playwright runs instead of Storybook. |
--cypress | Capture from Cypress runs instead of Storybook. |
--only-changed | Enable TurboSnap (snapshot only stories affected by changed files). |
--only-story-names | Run a build limited to specific story names. |
--auto-accept-changes | Auto-accept changes on the named branch (typically used on main for the first run after a planned UI overhaul). |
--exit-zero-on-changes | Exit 0 on visual changes (still uploads, but the CI step doesn't fail). |
--exit-once-uploaded | Return as soon as the upload completes; do not wait for the snapshot pass. |
--dry-run | Skip publishing; useful for local debugging. |
--no-interactive | CI-style logging when running locally. |
--trace-changed | Show the TurboSnap dependency tree (debug). |
--list | Print the indexed stories without running them. |
--debug | Verbose logging. |
--diagnostics-file <p> | Write a JSON diagnostics dump for support. |
Configuration file
chromatic.config.json at the project root - config keys mirror the CLI flags converted to kebab-case (cli (opens in new window)):
{
"$schema": "https://www.chromatic.com/config-file.schema.json",
"onlyChanged": true,
"exitZeroOnChanges": false,
"autoAcceptChanges": "main",
"externals": ["public/**", "tokens/**"]
}externals lists glob patterns of non-bundled files that should still trigger a TurboSnap rebuild when changed (e.g. design tokens, static assets) (turbosnap (opens in new window)).
Reading exit codes
Per the Chromatic CLI docs (opens in new window), exit codes carry semantic meaning - the gate logic should treat visual changes (1) differently from system errors (3, 21 - 23, 101 - 105, 201 - 220):
| Code | Meaning |
|---|---|
0 | Success. |
1 | Build has visual changes (review required). |
2 | Build has component (Storybook) errors. |
3 | Build failed (system error). |
4 | No stories found. |
5 - 6 | Build limited / canceled. |
11 - 12 | Account quota / payment issues. |
21 - 23 | Storybook build / load failures. |
101 - 105 | Git or dependency issues. |
201 - 220 | Network / API errors. |
254 - 255 | Invalid options / unknown error. |
Use --exit-zero-on-changes only when changes are reviewed in the Chromatic UI (not in CI) and you want CI to remain green pending review.
TurboSnap
TurboSnap analyzes git history and the bundler dependency graph to identify stories affected by a PR's changes; only those stories are snapshotted, and the rest copy their baseline forward (turbosnap (opens in new window)).
Enable via --only-changed (CLI) or onlyChanged: true (chromatic.config.json).
Bundler requirements
(Per turbosnap (opens in new window).)
Full-rebuild triggers
TurboSnap conservatively snapshots everything when any of these change (turbosnap (opens in new window)):
Merge commits also trigger conservative re-testing because TurboSnap cannot know ahead of time which baseline side to compare against (turbosnap (opens in new window)).
CI integration
The minimal pattern is: install chromatic, expose CHROMATIC_PROJECT_TOKEN as a CI secret, run the CLI, fail the job on exit != 0 (or != 0,1 if review happens in the Chromatic UI).
# .github/workflows/chromatic.yml
name: chromatic
on:
pull_request:
push:
branches: [main]
jobs:
chromatic:
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v5
with:
fetch-depth: 0 # required by TurboSnap
- uses: actions/setup-node@v4
with:
node-version: '20'
- run: npm ci
- name: Build Storybook + run Chromatic
env:
CHROMATIC_PROJECT_TOKEN: ${{ secrets.CHROMATIC_PROJECT_TOKEN }}
run: npx chromatic --only-changedfetch-depth: 0 is required so TurboSnap can resolve the PR-vs-baseline commit ancestry (turbosnap (opens in new window)).
References
Related skills
percy-visual-regression-testing
Authors Percy visual snapshot tests via the @percy/cli + framework SDK (Playwright, Cypress, Selenium, Storybook), runs them with `percy exec -- {test command}`, configures viewports / masking / ignored regions, and reviews diffs in the Percy build UI. Use when the project ships visual regression coverage to BrowserStack Percy.
playwright-snapshots
Authors Playwright `expect(page).toHaveScreenshot()` assertions, configures masks / clips / threshold / maxDiffPixels per test, manages the per-OS / per-browser snapshot directory, and runs the update flow with `--update-snapshots`. Use when the project ships self-hosted visual regression coverage in Playwright (no external snapshot service).
responsive-breakpoint-runner
Produces a single breakpoint-matrix report (rows = pages/stories, columns = viewports) across Percy, Chromatic, Playwright snapshots, or Storybook test-runner. Routes per-engine viewport syntax, runs each, and aggregates the results into one cross-breakpoint view. Use when the team needs one unified pass/fail view across three or more viewport widths instead of separate per-engine or per-breakpoint reports. The matrix-view output is the distinguishing trait: it dispatches to playwright-snapshots (and the other engines) rather than replacing any single-engine skill.
storybook-visual-regression-testing
Sets up visual regression coverage for a Storybook project - either via the official @chromatic-com/storybook addon (hosted) or via @storybook/test-runner with a postVisit hook that calls Playwright's toHaveScreenshot (self-hosted). Covers test-runner install, lifecycle hooks (setup / preVisit / postVisit), and CI integration. Use when a repo already has a working `.storybook/` config and the team wants per-story visual coverage rather than page-level snapshots.
visual-baseline-conventions
Reference catalog for visual regression coverage decisions - which Storybook stories or pages get baselines, how to choose breakpoints, when to mask vs adjust threshold, when to add or remove a baseline, and a decision matrix for picking among Percy / Chromatic / Playwright / Storybook test-runner. Use when designing visual coverage for a new project or auditing an existing baseline set.
visual-baseline-gate
Consumes pre-classified visual-diff JSON and a reviewer-signed acceptance log to produce a single go/no-go CI verdict for visual regression. Blocks when intentional baseline changes lack a non-author reviewer sign-off or when regressions are present, and emits a markdown + JSON artifact for the CI step. Use this skill when the gate's input is pre-classified diff data and the enforcement concern is reviewer approval, not when the goal is fanning out to multiple engines (use a multi-engine CI orchestrator for that).
visual-diff-summarizer
Triages a PR's visual regression diffs when there are too many changed screenshots to review one by one. Clusters snapshots (from Percy, Chromatic, Playwright `toHaveScreenshot`, Storybook, Loki) by component / route, separates changes that match PR intent from cascade / regression suspects, recommends which baselines to update, and emits one PR comment pointing the reviewer at the screenshots that need actual eyes. Use when a PR has 20+ visual diffs / changed screenshots and the reviewer needs help deciding which to open.