{
  "markdown": "# uxlint\n\nAudit any website's UX the way a design-literate reviewer would: contrast, tap targets, type\nscale, colour discipline, scan patterns, landmarks. Every finding comes with a prescriptive fix an\nagent (or a human) can apply directly. It's designed to sit in a coding agent's loop (MCP) and be\niterated against until green.\n\n![An agent audits a pricing page, gets a contrast error and a colour-clash finding with fixes for\neach, applies them, re-checks the rule, and re-grades the page from B to A](assets/demo.gif)\n\n*A real run, start to finish: `audit_url` → **Grade B**, a 2.39:1 contrast error and three CTAs in\nthree different accent hues → the fix → `verify_fix` → **Grade A**. Every number in it came back\nfrom the tools; only the waiting was cut.*\n\nThis is the **CLI**: a small, single static Rust binary. It drives a Chrome/Chromium you already\nhave installed over the DevTools protocol (no Node, no Playwright, no headless-browser download),\ncaptures what a page looks and reads like, and sends that to uxlint's hosted server, which does\nthe actual grading. The rules engine, the calibrated thresholds, and the LLM judge all live\nserver-side, so the client never needs updating when a rule changes.\n\n```\n┌──────────────────────────┐        POST /v1/audit {snapshots}        ┌──────────────────────────┐\n│ uxlint (this binary)     │ ───────────────────────────────────────▶ │ uxlint-server (hosted)    │\n│ drives YOUR Chrome (CDP) │ ◀─────────────────────────────────────── │ rules engine + LLM judge  │\n└──────────────────────────┘        report {findings + fixes}         └──────────────────────────┘\n```\n\n## Install\n\n```sh\ncurl -fsSL https://uxlint.net/install.sh | sh    # detects OS/arch, verifies checksum\n```\n\nOr with [mise](https://mise.jdx.dev) — its `github` backend pulls the matching build from GitHub\nReleases, verifies it, and updates on `mise up`:\n\n```sh\nmise use -g \"github:uxlint-net/uxlint-cli[rename_exe=uxlint]@latest\"\n```\n\nor pin it in a project's `mise.toml`:\n\n```toml\n[tools]\n\"github:uxlint-net/uxlint-cli\" = { version = \"latest\", rename_exe = \"uxlint\" }\n```\n\nOr build from source (needs a recent stable Rust toolchain and a Chrome/Chromium on PATH):\n\n```sh\ngit clone https://github.com/uxlint-net/uxlint-cli && cd uxlint-cli\ncargo build --release\n./target/release/uxlint --version\n```\n\n## Quickstart\n\n```sh\nuxlint auth login                                        # opens your browser, saves a token\nuxlint audit --base https://your-site.com --routes /,/pricing\n```\n\nFirst time auditing your own project? `uxlint init` picks (or creates) a site to attach reports\nto and writes a `uxlint.toml` so every future audit in this directory just works:\n\n```sh\nuxlint init\nuxlint audit --base http://localhost:5173 --routes /,/pricing\n```\n\nExit code 1 on findings above the configured severity → drop it straight into CI (see\n`.github/workflows/` for a template, or the `uxlint-net/uxlint-action` GitHub Action).\n\n## Hiding elements from an audit (`uxlint-hide`)\n\nSome on-page chrome isn't product UI and shouldn't be judged: a dev/staging environment banner, a\n\"DEV\" marker, a debug toolbar, a Storybook/preview affordance. Add the class **`uxlint-hide`** to any\nsuch element and the audit removes it — it's `display:none` from first paint, so it never appears in a\nscreenshot and is invisible to the collector (it seeds no findings):\n\n```html\n<div class=\"env-banner uxlint-hide\">STAGING</div>\n```\n\nThe class is **inert on your real site** — it does nothing unless the audit is running, because the\nstylesheet that hides it (`.uxlint-hide { display: none !important; }`) is injected only by uxlint's\nbrowser, before the page's own scripts run. Style your element however you like the rest of the time.\nIt applies in every capture path — the crawl, goal-walk tests, and fix previews.\n\n## MCP (use it from a coding agent)\n\n**Claude Code, one command:**\n\n```\n/plugin marketplace add uxlint-net/uxlint-cli\n/plugin install uxlint@uxlint\n```\n\nThat installs the `uxlint` MCP server and, if the CLI isn't already on your PATH, fetches the matching\nversion once with the same checksum-verifying installer as above — so `/plugin update` updates the CLI\nunderneath it too. **No Node needed**: it downloads one static binary (verified against a published\nchecksum) and drives the Chrome you already have.\n\n**Any other agent** — one line (the npm package fetches the binary for your platform, verifies the\nchecksum published beside it, and hands over). This is the only route that needs **Node 18+**, for\n`npx` itself; if you'd rather not, install the binary with the line at the top and register that:\n\n```sh\nclaude mcp add uxlint -- npx -y @uxlint-net/uxlint mcp\n```\n\nOr, for a client that reads a JSON config:\n\n```json\n{ \"mcpServers\": { \"uxlint\": { \"command\": \"npx\", \"args\": [\"-y\", \"@uxlint-net/uxlint\", \"mcp\"] } } }\n```\n\nuxlint is also in the [MCP Registry](https://registry.modelcontextprotocol.io) as\n`io.github.uxlint-net/uxlint`, for clients that browse it. Already have the CLI? `uxlint mcp install`\nregisters it directly, no npx wrapper.\n\nThere is no token to set up first: ask your agent to audit something while signed out and it hands you\na sign-in link that mints and saves the token for you (`UXLINT_API_KEY` is for CI, which has no\nbrowser).\n\nFive tools: `audit_url` (full audit, graded verdict + action plan), `verify_fix` (recheck one rule\non one page after an edit), `get_shot` (fetch a finding's annotated screenshot),\n`ux_guidance` (best-practice guidance to read *before* building UI), and `lint_feedback` — opt-in\nand off by default (§ Privacy) — one tool for three kinds of signal: whether a finding was useful,\na lint uxlint is missing, or a component library it didn't recognise. The agent audits, reads the\nfixes, edits, and re-audits until green.\n\n## Privacy & trust\n\nThis CLI runs on your machine and drives a real browser against real pages, so it's fair to ask\nexactly what it captures and where it goes. What we can tell you, because it's what the code in\nthis repo actually does:\n\n- **The collector is baked in and readable.** It's compiled into this\n  binary (`include_str!` of `assets/collector.js`), so `uxlint --version` pins the exact capture\n  code and the server can't inject anything at run time. Everything it captures is page geometry,\n  visible text, computed styles, and screenshots. For an embedded `<iframe>` it records the src's\n  **host** only — never the full embed URL, which can carry session ids and tokens in its query\n  string. It never reads your source code or your filesystem beyond `uxlint.toml`. It does read a\n  little **project provenance** and send it with the report:\n  your current git commit sha and branch name (`git rev-parse`), the machine's hostname, and, in\n  GitHub Actions, the repo/PR/commit link. Set `UXLINT_RUNNER` to override the hostname.\n- **Secret & PII redaction is best-effort, not a guarantee.** Before anything is uploaded, the\n  collector masks text that *looks like* a token, API key, password, or email address in captured\n  page text, and redacts the same patterns from console logs and native dialog messages. All\n  channels share one pattern list (`assets/redact.js`), so they can't drift. Screenshots get an\n  extra pass right before capture: every form field value is masked (passwords blanked, other\n  inputs replaced with dots) and pattern-matched secrets in on-page text are scrubbed, so typed data\n  and displayed keys don't land in the image. That pass reaches into shadow DOM (including closed\n  roots, via an `attachShadow` interceptor) and same-origin iframes, and covers a cross-origin\n  iframe with an opaque box since its pixels can't be redacted. But redaction is pattern-based, and\n  a screenshot is still pixels: arbitrary displayed content that no pattern catches (a customer name\n  on the page, order data), split-up values, and anything drawn into images or `<canvas>` can still\n  slip through. Credentials you pass with\n  `--header`/`--storage`/`--login-*` drive *your* browser only and are never sent to uxlint's server.\n\n  > **Because a report captures page HTML, text, and screenshots, it is impossible to fully guard\n  > against sensitive content leaking into it. Use TEST accounts, not real or production ones.** For\n  > local development the risk is low, as long as the data is only local development data. When you\n  > audit an authenticated site that holds real secrets or personal data, review what gets sent\n  > before you send it: use `--dry-run` to write the exact payload (page text, provenance, and\n  > screenshots) to a local folder and inspect it without uploading. Redaction reduces accidental\n  > exposure; it is not a security boundary, and you remain responsible for what you point uxlint at.\n- **Navigational text is scrubbed for secrets only, on purpose.** Control labels, menu and\n  `<select>` options, and workspace/org switcher names run through the same secret patterns, but they\n  are not redacted for names or other arbitrary content. The reason is the goal walk: an audit drives\n  the page with an LLM that reads exactly this text to find the right control, operate it, and match\n  its choice back to the DOM. Masking it would defeat the walk, because the judge could no longer\n  tell two options apart or click the one it picked. So the labels an audit needs to navigate stay\n  readable, and a real name that rides along in one of them is covered by the test-accounts rule\n  above rather than by redaction. This is a deliberate trade: keeping the goal walk working is worth\n  more than blanking text the test-accounts rule already protects.\n- **No telemetry.** This binary makes outbound calls to exactly the hosts you tell it to: the\n  uxlint API server (`--server`/`UXLINT_SERVER`, or the default hosted origin), the site you ask it\n  to audit, and, only if you explicitly opt in, anonymous rule-feedback signals. There is no\n  separate analytics/crash-reporting/phone-home destination baked in anywhere.\n- **Feedback is opt-in, off by default.** `uxlint init` asks once; it only ever writes\n  `feedback = true` to `uxlint.toml` if you say yes, and you can flip it back at any time.\n- **The audit browser uses an ephemeral profile.** Each audit launches Chrome with a fresh, throwaway\n  user-data directory, so no cookies, history, or extensions from your everyday browsing are ever\n  loaded into the audited session, and nothing persists after the process exits.\n- **Your login stays local.** `uxlint auth login` stores a token at\n  `~/.config/uxlint/credentials`, chmod'd `0600`. It's never logged, never printed (except the one\n  deliberate case: `uxlint signup` prints a freshly minted key so you can export it), and never\n  bundled into a report.\n\nThis isn't a substitute for reading the source. It's short, and that's rather the point of\npublishing it. If you find something that doesn't match this description, please open an issue.\n\n## What this CLI is *not*\n\nIt's deliberately dumb: navigate, run the baked-in collector, upload the snapshot, print the\nreport. The rules, thresholds, and judge model are not in this repo and never will be. They're\nthe actual product, and they live server-side only. A build of this CLI is useless without a\nuxlint server to talk to (the hosted one at `https://uxlint.net` by default, or your own).\n\n## License\n\nApache License 2.0 (see `LICENSE`). Read it, audit it, fork it, build it from source, ship it\ninside your own tooling — no conditions beyond the usual attribution and patent terms.\n\nThis used to be the Business Source License, which converted to Apache-2.0 on a Change Date in\n2030; we've simply arrived early. The restriction it carried — no competing hosted \"audit my site\"\nservice built on this code — was protecting the wrong thing: what's valuable is the rules, the\ncalibrated thresholds and the judge, and those are server-side and not in this repo. What IS here is\na client that needs a uxlint server to be worth anything, and a client is exactly the piece that\nshould be frictionless to install, read and vendor.\n\nReleases up to and including **v0.1.30** were published under BUSL-1.1; **v0.1.31 onward** are\nApache-2.0.\n",
  "bytes": 12143,
  "sha": "91b6f68ad10c07442a16adb442669ba495d01c9777ea5c86802b4d3f416016f6",
  "repo_slug": "uxlint-net/uxlint-cli",
  "fonte": "repo",
  "truncated": false,
  "api": "https://agentalog.com/api/listings/mcp_io_github_uxlint_net_uxlint_fa7c5351/readme"
}