sniff-qa 0.6.1 → 0.8.0

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (71) hide show
  1. package/.claude-plugin/plugin.json +2 -2
  2. package/CHANGELOG.md +96 -0
  3. package/NOTICE +1 -1
  4. package/README.md +109 -320
  5. package/dist/assert-budget-NBF3JYM7.js +9 -0
  6. package/dist/caps-G3ZGJI7B.js +21 -0
  7. package/dist/caps-G3ZGJI7B.js.map +1 -0
  8. package/dist/{chunk-LJYT4MD5.js → chunk-7ZOFYET5.js} +18 -97
  9. package/dist/chunk-7ZOFYET5.js.map +1 -0
  10. package/dist/chunk-EGANHIIE.js +91 -0
  11. package/dist/chunk-EGANHIIE.js.map +1 -0
  12. package/dist/chunk-GV27IOMC.js +106 -0
  13. package/dist/chunk-GV27IOMC.js.map +1 -0
  14. package/dist/chunk-GX36BPK7.js +161 -0
  15. package/dist/chunk-GX36BPK7.js.map +1 -0
  16. package/dist/chunk-NG2IGCAK.js +45 -0
  17. package/dist/chunk-NG2IGCAK.js.map +1 -0
  18. package/dist/cli/index.js +27 -16
  19. package/dist/cli/index.js.map +1 -1
  20. package/dist/crawl-ZXZRF4Q7.js +15 -0
  21. package/dist/crawl-ZXZRF4Q7.js.map +1 -0
  22. package/dist/crawl-command-RDWGHZC4.js +76 -0
  23. package/dist/crawl-command-RDWGHZC4.js.map +1 -0
  24. package/dist/{discover-Q363YJ5X.js → discover-QKGMYPHK.js} +15 -7
  25. package/dist/discover-QKGMYPHK.js.map +1 -0
  26. package/dist/{domain-NNL2PPRK.js → domain-UQ34YE4B.js} +1 -1
  27. package/dist/index.d.ts +15 -0
  28. package/dist/index.js +6 -2
  29. package/dist/index.js.map +1 -1
  30. package/dist/{loader-UG2XADG7.js → loader-Z2DUGG5Z.js} +18 -2
  31. package/dist/loader-Z2DUGG5Z.js.map +1 -0
  32. package/dist/mcp/server.d.ts +20 -1
  33. package/dist/mcp/server.js +37 -28
  34. package/dist/mcp/server.js.map +1 -1
  35. package/dist/{model-2DTSAX7T.js → model-SP6HJN6J.js} +4 -2
  36. package/dist/model-SP6HJN6J.js.map +1 -0
  37. package/dist/{persistence-RHUFBQAY.js → persistence-XLFJ5EY4.js} +6 -2
  38. package/dist/persistence-XLFJ5EY4.js.map +1 -0
  39. package/dist/{quality-run-LLBZZI7X.js → quality-run-YAURB242.js} +6 -5
  40. package/dist/{quality-run-LLBZZI7X.js.map → quality-run-YAURB242.js.map} +1 -1
  41. package/dist/redaction-CHAIVPCF.js +30 -0
  42. package/dist/redaction-CHAIVPCF.js.map +1 -0
  43. package/dist/{repo-analyzer-WK5WP7NT.js → repo-analyzer-4OURSITY.js} +1 -1
  44. package/dist/{report-IKRVVIW4.js → report-7FQFH4HW.js} +3 -3
  45. package/dist/{runner-HFJDHS2I.js → runner-QHSYX7J2.js} +1 -1
  46. package/dist/{runner-II7BPB63.js → runner-RDVVIGBU.js} +7 -2
  47. package/dist/runner-RDVVIGBU.js.map +1 -0
  48. package/dist/{unified-67Z6YA64.js → unified-YSOTPG5B.js} +12 -10
  49. package/dist/{unified-67Z6YA64.js.map → unified-YSOTPG5B.js.map} +1 -1
  50. package/dist/{update-baselines-UCAW3SDN.js → update-baselines-JYKUUIZG.js} +2 -2
  51. package/dist/write-report-6RJWN54I.js +31 -0
  52. package/dist/write-report-6RJWN54I.js.map +1 -0
  53. package/package.json +5 -4
  54. package/skills/sniff/SKILL.md +10 -10
  55. package/skills/sniff-fix/SKILL.md +3 -0
  56. package/skills/sniff-report/SKILL.md +3 -0
  57. package/dist/chunk-LJYT4MD5.js.map +0 -1
  58. package/dist/crawl-666NLYDC.js +0 -11
  59. package/dist/crawl-command-3OB4IUBU.js +0 -54
  60. package/dist/crawl-command-3OB4IUBU.js.map +0 -1
  61. package/dist/discover-Q363YJ5X.js.map +0 -1
  62. package/dist/loader-UG2XADG7.js.map +0 -1
  63. package/dist/model-2DTSAX7T.js.map +0 -1
  64. package/dist/persistence-RHUFBQAY.js.map +0 -1
  65. package/dist/runner-II7BPB63.js.map +0 -1
  66. /package/dist/{crawl-666NLYDC.js.map → assert-budget-NBF3JYM7.js.map} +0 -0
  67. /package/dist/{domain-NNL2PPRK.js.map → domain-UQ34YE4B.js.map} +0 -0
  68. /package/dist/{repo-analyzer-WK5WP7NT.js.map → repo-analyzer-4OURSITY.js.map} +0 -0
  69. /package/dist/{report-IKRVVIW4.js.map → report-7FQFH4HW.js.map} +0 -0
  70. /package/dist/{runner-HFJDHS2I.js.map → runner-QHSYX7J2.js.map} +0 -0
  71. /package/dist/{update-baselines-UCAW3SDN.js.map → update-baselines-JYKUUIZG.js.map} +0 -0
@@ -1,9 +1,9 @@
1
1
  {
2
2
  "name": "sniff",
3
- "version": "0.6.1",
3
+ "version": "0.8.0",
4
4
  "description": "Autonomous QA scanner via sniff-qa: walks your running app in a real browser and reports real bugs (broken pages/links, console/network errors, broken forms, state-loss, empty/placeholder data, bad loading/error states, responsive + a11y) with reproduction proof, severity, confidence, and a fix. No API key.",
5
5
  "author": {
6
- "name": "Adam Boudj",
6
+ "name": "Adam Boudjemaa",
7
7
  "url": "https://github.com/Aboudjem/sniff"
8
8
  },
9
9
  "homepage": "https://github.com/Aboudjem/sniff",
package/CHANGELOG.md CHANGED
@@ -7,6 +7,102 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
7
7
 
8
8
  ## [Unreleased]
9
9
 
10
+ ## [0.8.0] - 2026-09-02
11
+
12
+ Three capability additions (MCP tool gating, authenticated walks, assertion budgets), a Neon Noir
13
+ visual identity, a rewritten README with four refreshed translations, and an installer that
14
+ delegates to the Vercel skills CLI. The scanner's findings and their wording are unchanged.
15
+
16
+ ### Added
17
+
18
+ - **`--caps` for the MCP server.** Five capabilities (`scan`, `walk`, `discover`, `report`,
19
+ `install`) narrow the registered tool set and the unified `sniff` tool's `mode` enum. The default
20
+ grants all five, so an existing config keeps its exact tool list and registration order. Unknown or
21
+ empty lists exit 1 on stderr. Logic lives in `src/mcp/caps.ts`; 14 tests, three of which drive a
22
+ real MCP client over `InMemoryTransport` and assert the actual `tools/list` result.
23
+ - **`--storage-state <path>` for walking a logged-in app.** Threaded through `browserConfigSchema`,
24
+ the CLI flag, both crawl contexts, `BrowserRunner`, and the MCP walk. Cookie values, localStorage
25
+ values, JSON string leaves inside them, indexedDB strings, percent-encoded forms, and
26
+ `Authorization` header values are redacted from progress output, screenshot filenames, the returned
27
+ report, `saveResults`, and `saveReport`. Screenshot pixels are out of scope and stay unredacted.
28
+ 21 tests, including one that plants four tokens across a report and asserts each appears zero times
29
+ in the written HTML and JSON.
30
+ - **An `assert` block in `sniff.config`.** `maxCritical`, `maxHigh`, `maxMedium`, `maxLow`,
31
+ `maxInfo`, and `maxTotal` cap findings by severity. Additive only: it can turn a pass into a
32
+ failure, never the reverse, and an absent block behaves exactly as before. Enforced on the command
33
+ line by the walk, the public `sniff scan`, the browser mode, and discovery. 31 tests.
34
+ - **Neon Noir visual identity.** New `hero-dark.svg` and `hero-light.svg` for the README first
35
+ screen, `logo-mark.png` and `logo-mark-512.png` as a square mark, and a rebuilt 1280x640
36
+ `social-preview.png`. `demo.svg`, `features-dark.svg`, `features-light.svg`, `how-it-works.svg`,
37
+ `logo-dark.svg`, `logo-light.svg`, and `social-preview.svg` were rebuilt in the new palette.
38
+ - **New docs**: `docs/editors.md`, `docs/authenticated-walks.md`, `docs/assert-budgets.md`,
39
+ `docs/comparison.md`, and `docs/faq.md`.
40
+
41
+ ### Changed
42
+
43
+ - **`install.sh` delegates to the Vercel skills CLI** by default
44
+ (`npx --yes skills@1.5.23 add Aboudjem/sniff -a <agent> -y`). All 13 platform ids map to agent
45
+ codes that exist in the live supported-agents table, which also corrects four skills directories
46
+ the hand-maintained table had wrong. The clone-and-symlink path stays reachable as `--legacy` and
47
+ is selected automatically when `npx` is missing. `--update`, `--uninstall`, and `--no-mcp` work on
48
+ both paths.
49
+ - **README rewritten** from 452 lines to 180, to the plugin skeleton: one install block above the
50
+ first heading, one editor table in place of an eight-row matrix and seven per-editor `<details>`
51
+ blocks, and every image path absolute so it renders on npmjs.com. The comparison table, the metrics
52
+ history, and the long FAQ moved to `docs/comparison.md` and `docs/faq.md`.
53
+ - **All four translations** (`READMEs/zh-CN.md`, `ja.md`, `es.md`, `fr.md`) rewritten from the new
54
+ English source.
55
+ - **Plugin manifests** (`.claude-plugin`, `.cursor-plugin`, `.copilot-plugin`), `server.json`,
56
+ `package.json`, and `package-lock.json` all moved to 0.8.0 together. Previously `package.json` and
57
+ `server.json` sat at 0.7.1 while the three manifests read 0.7.0.
58
+ - **`skills/sniff-fix/SKILL.md` and `skills/sniff-report/SKILL.md`** now say in their bodies that
59
+ they are invoked by the `sniff` skill's own flow, that `user-invocable: false` is honored only by
60
+ Claude Code, and that other agents may surface them as standalone commands.
61
+
62
+ ### Removed
63
+
64
+ - `.github/assets/sniff-diagram.svg`. `how-it-works.svg` draws the same five-step flow, and its name
65
+ was dropped from the copy loop in `.github/workflows/deploy-pages.yml`.
66
+ - The Star History block, the three social badges, and the built-on credit list from the README
67
+ footer.
68
+
69
+ ### Fixed
70
+
71
+ - `READMEs/fr.md`: the French translation carried injected placeholder text and a corrupted install
72
+ command from an earlier machine-translation pass. All four translations are now generated from the
73
+ current English README, with code blocks, asset URLs, and commands verbatim.
74
+ - Em-dashes removed from `package.json`, `src/mcp/server.ts`, `src/mcp/handlers.ts`,
75
+ `src/cli/index.ts`, `src/cli/commands/crawl-command.ts`, `src/cli/commands/discover.ts`, and
76
+ `src/report/model.ts`.
77
+
78
+ ## [0.7.0] - 2026-05-29
79
+
80
+ Portability and discoverability layer: a multi-CLI installer, auto-discovery manifests, a GitHub Pages demo, and localized READMEs, on top of the earlier documentation polish. No change to the scanner engine or its findings.
81
+
82
+ ### Added
83
+
84
+ - Multi-CLI installer: `install.sh` and `install.ps1` symlink the `/sniff`, `/sniff-fix`, and `/sniff-report` skills into a target CLI's skills directory, with `--update` and `--uninstall` and a one-line curl install. The MCP server stays the universal fallback.
85
+ - Auto-discovery manifests: `.cursor-plugin/plugin.json` and `.copilot-plugin/plugin.json`, mirroring `.claude-plugin/plugin.json`.
86
+ - GitHub Pages demo: `site/index.html` plus `.github/workflows/deploy-pages.yml`, reusing the existing demo assets and a sample report.
87
+ - README: a language-switcher row, a multi-platform install matrix with a one-line curl, an npm downloads badge, and a Star History chart.
88
+ - Localized READMEs under `READMEs/` for zh-CN, ja, es, and fr (machine-assisted; English stays the source of truth).
89
+ - Demo video: `.github/assets/demo.gif` (hero) plus an `.mp4` and an embed guide.
90
+
91
+ ### Changed
92
+
93
+ - Rebuilt the features and hero SVGs around the flow-walk model, and removed emoji from them.
94
+ - Named the slash commands and MCP tools in the README.
95
+
96
+ ### Removed
97
+
98
+ - Stale pre-0.6.0 SVG assets and the internal rebuild docs.
99
+
100
+ ### Fixed
101
+
102
+ - `.github/FUNDING.yml`: corrected the GitHub Sponsors handle to `Aboudjem` (the previous value did not resolve).
103
+ - Animated SVGs now respect `prefers-reduced-motion`.
104
+ - Removed sentence-break em-dashes from the docs.
105
+
10
106
  ## [0.6.1] - 2026-05-28
11
107
 
12
108
  Alignment release: no behaviour changes. Fixes the release pipeline so future tags can publish to npm,
package/NOTICE CHANGED
@@ -5,7 +5,7 @@ This product includes software developed by Adam Boudj.
5
5
 
6
6
  If you use Sniff in your project, a mention or attribution is appreciated:
7
7
  - In your README, docs, or project credits
8
- - "QA powered by Sniff (https://github.com/adamboudj/sniff)"
8
+ - "QA powered by Sniff (https://github.com/Aboudjem/sniff)"
9
9
 
10
10
  Third-party licenses:
11
11
  - Playwright (Apache-2.0) - https://playwright.dev
package/README.md CHANGED
@@ -1,392 +1,181 @@
1
1
  <picture>
2
- <source media="(prefers-color-scheme: dark)" srcset="https://raw.githubusercontent.com/Aboudjem/sniff/main/.github/assets/logo-dark.svg">
3
- <source media="(prefers-color-scheme: light)" srcset="https://raw.githubusercontent.com/Aboudjem/sniff/main/.github/assets/logo-light.svg">
4
- <img alt="Sniff" src="https://raw.githubusercontent.com/Aboudjem/sniff/main/.github/assets/logo-light.svg" width="100%">
2
+ <source media="(prefers-color-scheme: dark)" srcset="https://raw.githubusercontent.com/Aboudjem/sniff/main/.github/assets/hero-dark.svg">
3
+ <source media="(prefers-color-scheme: light)" srcset="https://raw.githubusercontent.com/Aboudjem/sniff/main/.github/assets/hero-light.svg">
4
+ <img alt="sniff" src="https://raw.githubusercontent.com/Aboudjem/sniff/main/.github/assets/hero-light.svg" width="100%">
5
5
  </picture>
6
6
 
7
7
  <p align="center">
8
- <a href="https://www.npmjs.com/package/sniff-qa"><img src="https://img.shields.io/npm/v/sniff-qa?color=ef4444&logo=npm&label=npm&style=flat-square" alt="npm"></a>
9
- <a href="LICENSE"><img src="https://img.shields.io/badge/license-Apache_2.0-ef4444?style=flat-square" alt="License"></a>
10
- <a href="https://github.com/Aboudjem/sniff/actions/workflows/ci.yml"><img src="https://img.shields.io/github/actions/workflow/status/Aboudjem/sniff/ci.yml?style=flat-square&label=CI" alt="CI"></a>
11
- <a href="https://nodejs.org"><img src="https://img.shields.io/badge/node-%E2%89%A522-339933?style=flat-square&logo=node.js&logoColor=white" alt="Node"></a>
12
- <a href="https://github.com/Aboudjem/10x"><img src="https://img.shields.io/badge/10x-marketplace-ef4444?style=flat-square" alt="10x marketplace"></a>
13
- <a href="https://github.com/Aboudjem/sniff/stargazers"><img src="https://img.shields.io/github/stars/Aboudjem/sniff?style=flat-square&color=ef4444" alt="Stars"></a>
8
+ <a href="https://www.npmjs.com/package/sniff-qa"><img src="https://img.shields.io/npm/v/sniff-qa?style=flat-square&color=FF006E&logo=npm&label=npm" alt="npm version"></a>
9
+ <a href="https://github.com/Aboudjem/sniff/actions/workflows/ci.yml"><img src="https://img.shields.io/github/actions/workflow/status/Aboudjem/sniff/ci.yml?style=flat-square&color=00D4FF&label=CI" alt="CI"></a>
10
+ <a href="LICENSE"><img src="https://img.shields.io/badge/license-Apache_2.0-7C3AED?style=flat-square" alt="License Apache 2.0"></a>
11
+ <a href="https://github.com/Aboudjem/sniff/stargazers"><img src="https://img.shields.io/github/stars/Aboudjem/sniff?style=flat-square&color=2BE8C8" alt="Stars"></a>
14
12
  </p>
15
13
 
16
- <p align="center"><b>Point it at your running app. It walks your real user flows in a real browser and tells you what's actually broken — with proof.</b></p>
14
+ <p align="center"><b>English</b> · <a href="READMEs/zh-CN.md">简体中文</a> · <a href="READMEs/ja.md">日本語</a> · <a href="READMEs/es.md">Español</a> · <a href="READMEs/fr.md">Français</a></p>
17
15
 
18
- <p align="center">
19
- <a href="#get-started">Get started</a> ·
20
- <a href="#what-it-finds">What it finds</a> ·
21
- <a href="#why-you-can-trust-it">Why trust it</a> ·
22
- <a href="#how-it-works">How it works</a> ·
23
- <a href="#demo">Demo</a> ·
24
- <a href="#faq">FAQ</a>
25
- </p>
26
-
27
- <picture>
28
- <source media="(prefers-color-scheme: dark)" srcset="https://raw.githubusercontent.com/Aboudjem/sniff/main/.github/assets/sniff-diagram.svg">
29
- <img alt="Sniff flow: your running app -> headless browser walk -> findings with proof" src="https://raw.githubusercontent.com/Aboudjem/sniff/main/.github/assets/sniff-diagram.svg" width="100%">
30
- </picture>
31
-
32
- ---
33
-
34
- ## What is this?
35
-
36
- Sniff is an autonomous QA scanner. You point it at a running web app and it **walks your app's real user flows in a real (headless) browser** — clicking buttons, filling forms, following links — and reports what's actually broken.
16
+ <p align="center"><b>Point it at your running app. It walks your real user flows in a real browser and tells you what is actually broken, with proof.</b></p>
37
17
 
38
- It is **not** a linter and **not** a static scanner. It opens your pages, interacts with them like a user would, and watches what happens.
18
+ <p align="center"><a href="#what-it-does">What it does</a> · <a href="#install">Install</a> · <a href="#use-it">Use it</a> · <a href="#what-you-get">What you get</a> · <a href="#works-in-your-editor">Works in your editor</a> · <a href="#good-to-know">Good to know</a></p>
39
19
 
40
- Every bug it reports comes with **proof**: the exact page, the ordered steps to reproduce it, a screenshot, and the console or network excerpt that caught it. A finding without reproduction proof is not a finding.
20
+ <img alt="sniff walking a buggy app and streaming findings with severity, confidence, steps to reproduce, and a fix" src="https://raw.githubusercontent.com/Aboudjem/sniff/main/.github/assets/demo.gif" width="100%">
41
21
 
42
22
  ```bash
43
- npx sniff-qa --url http://localhost:3000
23
+ claude plugin marketplace add Aboudjem/10x
24
+ claude plugin install sniff@10x
44
25
  ```
45
26
 
46
- One command — no config, no API key, no Playwright setup. Point it at your running app, or just run `npx sniff-qa` from your project and it auto-detects a dev server on common ports.
27
+ ## What it does
47
28
 
48
- > Sniff walks a **running app**. If no dev server is running, it falls back to a source-code scan and tells you exactly how to start the real walk. See [Get started](#get-started).
49
-
50
- ---
29
+ Most linters read your source without ever running your app, and end-to-end frameworks ask you to
30
+ write and maintain the tests yourself. sniff opens your running app in a real browser, clicks and
31
+ fills things like a user would, and judges what actually happened.
51
32
 
52
- ## Get started
33
+ - **It finds 12 classes of bugs**, from HTTP 500 routes and dead links to placeholder data, dead
34
+ submit buttons, forms wiped by the back button, stuck spinners, and mobile overflow.
35
+ - **It proves every one.** Every finding carries the route and the ordered steps that produced it,
36
+ plus the screenshot and the console or network excerpt the check captured. No steps, no finding.
37
+ - **It has been measured.** On a fixture app planted with 21 bugs across all 12 classes, plus a
38
+ clean control page, sniff finds 21 of 21 and reports nothing on the control page.
53
39
 
54
- You need **Node.js 22+** and a web app you can run locally (or a URL).
40
+ ## Install
55
41
 
56
- > **Naming:** the npm package is **`sniff-qa`** — use `npx sniff-qa` or `npm install -D sniff-qa`. Once installed, the binary is **`sniff`** (the `sniff-qa` binary works too). Don't run `npx sniff` — that's a different package.
57
-
58
- ### 1. Start your app
42
+ The block above is the Claude Code path, through the [10x marketplace](https://github.com/Aboudjem/10x).
43
+ For any other agent, the Vercel skills CLI installs the same three skills:
59
44
 
60
45
  ```bash
61
- npm run dev # or however you start your app
46
+ npx skills add Aboudjem/sniff
62
47
  ```
63
48
 
64
- ### 2. Walk it
65
-
66
- In another terminal:
49
+ To use it as a plain command line tool, with no editor involved:
67
50
 
68
51
  ```bash
69
- npx sniff-qa --url http://localhost:3000 # point it at your running app
52
+ npx sniff-qa --url http://localhost:3000
70
53
  ```
71
54
 
72
- That's the reliable one-liner. Sniff also **auto-detects** a dev server on common ports, so from your project folder you can often just run `npx sniff-qa` with no flags. If your app is on a non-standard port (or auto-detect misses it), pass `--url` — that always works. It also walks a deployed URL:
55
+ The npm package is `sniff-qa` and the binary it installs is `sniff`. Do not run `npx sniff`, which
56
+ is an unrelated package.
73
57
 
74
- ```bash
75
- npx sniff-qa --url https://staging.myapp.com
76
- ```
77
-
78
- > **Found bugs? It exits non-zero — on purpose.** A walk that finds issues exits with code `1` so CI fails the build; it is **not** a crash (you'll see a `✓ Scan complete` line). Pass `--fail-on none` to always exit `0`.
58
+ <details>
59
+ <summary>Node version, project install, and CI</summary>
79
60
 
80
- > **First run downloads a browser.** The first time Sniff opens a browser it downloads a Chromium build (~165 MB, one-time). You'll see the progress. You need internet access for that first run; after that it's cached.
61
+ Node.js 22 or newer. `npm install -D sniff-qa` adds it to a project's devDependencies, and
62
+ `npx sniff-qa ci` writes a GitHub Actions workflow with browser caching and report artifacts.
63
+ </details>
81
64
 
82
- ### 3. Read the report
65
+ ## Use it
83
66
 
84
- Findings print to your terminal, grouped by severity, each with steps to reproduce. Want a shareable page?
67
+ **1. Start your app,** with whatever dev server your project already uses.
85
68
 
86
69
  ```bash
87
- npx sniff-qa --url http://localhost:3000 --report # runs a walk, then writes sniff-reports/sniff-report.html (self-contained)
70
+ npm run dev
88
71
  ```
89
72
 
90
- **No app running?** Sniff doesn't fail silently. It runs a source-only scan and prints a clear next step — start your dev server or pass `--url` — so you can get to the real flow-walk. You can also run the source scan on purpose:
73
+ **2. Walk it,** from a second terminal. sniff auto-detects a dev server on the common ports, so
74
+ `--url` is optional, but passing it takes the guesswork out.
91
75
 
92
76
  ```bash
93
- npx sniff-qa scan # source-only scan, no browser
77
+ npx sniff-qa --url http://localhost:3000
94
78
  ```
95
79
 
96
- Stuck? Run `npx sniff-qa doctor` to check your environment (Node, browser, dev server).
80
+ **3. Read the findings.** They print grouped by severity. Below is an abridged real run against this
81
+ repo's own planted-bug fixture, from `npx sniff-qa --url http://localhost:4321 --ci --max-pages 12`:
97
82
 
98
- ---
83
+ ```text
84
+ sniff v0.8.0 walking http://localhost:4321
85
+
86
+ 26 findings (+1 low-confidence hidden; use --all)
99
87
 
100
- ## What it finds
88
+ CRITICAL (1)
89
+ • [confirmed] Page returns HTTP 500
90
+ /crash (route/broken-page)
91
+ - Navigate to /crash
92
+ - Server responded with HTTP 500
93
+ fix: The route throws server-side. Check the server logs/handler for this path and return a valid page or a proper error page.
94
+ shot: sniff-reports/crawl/_crash-desktop.png
101
95
 
102
- Sniff walks your app and looks for **12 classes of real bugs**:
96
+ ✓ Scan complete: 26 issue(s) found. Exit code 1 so CI fails on bugs; pass --fail-on none to always exit 0.
97
+ ```
103
98
 
104
- | # | Class | Examples |
105
- |:--|:------|:---------|
106
- | 1 | **Broken pages / routes** | 4xx/5xx responses, blank renders, crash screens |
107
- | 2 | **Broken links** | Dead internal and external links |
108
- | 3 | **Console & network errors** | Uncaught exceptions and failed requests *during interaction* |
109
- | 4 | **Empty & fake data** | Missing data, plus placeholders like `lorem ipsum`, `TODO`, `test@test.com` |
110
- | 5 | **Broken forms** | Dead submit buttons, validation that never fires |
111
- | 6 | **State loss** | Fill a form, hit back, and it's wiped |
112
- | 7 | **Flow regressions / dead-ends** | A journey that can't be completed |
113
- | 8 | **Bad loading & error states** | Infinite spinners, missing error states |
114
- | 9 | **Broken async outcomes** | Submitted but no success feedback (flagged "needs out-of-band verification") |
115
- | 10 | **Responsive issues** | Overflow and tiny tap targets (a 375px mobile pass) |
116
- | 11 | **Accessibility** | Missing alt text and labels, contrast — via [axe-core](https://github.com/dequelabs/axe-core) |
117
- | 12 | **Unclear primary actions** | The main call-to-action is buried or ambiguous |
99
+ Add `--report` for a self-contained HTML page you can send to someone. Run `npx sniff-qa doctor` if
100
+ the environment looks wrong.
118
101
 
119
- Each finding ships with:
102
+ <img alt="How sniff works: crawl, act, assert, prove, report" src="https://raw.githubusercontent.com/Aboudjem/sniff/main/.github/assets/how-it-works.svg" width="100%">
120
103
 
121
- - **Reproduction proof** — the exact route, the ordered steps, a screenshot, and the console/network excerpt.
122
- - **A severity** — so you fix the right thing first.
123
- - **A confidence** — `confirmed`, `likely`, or `uncertain`. Uncertain findings are hidden by default; add `--all` to see them.
124
- - **A suggested fix.**
104
+ ## What you get
125
105
 
126
106
  <picture>
127
107
  <source media="(prefers-color-scheme: dark)" srcset="https://raw.githubusercontent.com/Aboudjem/sniff/main/.github/assets/features-dark.svg">
128
108
  <source media="(prefers-color-scheme: light)" srcset="https://raw.githubusercontent.com/Aboudjem/sniff/main/.github/assets/features-light.svg">
129
- <img alt="The bug classes Sniff finds" src="https://raw.githubusercontent.com/Aboudjem/sniff/main/.github/assets/features-light.svg" width="100%">
109
+ <img alt="The 12 classes of bugs sniff finds" src="https://raw.githubusercontent.com/Aboudjem/sniff/main/.github/assets/features-light.svg" width="100%">
130
110
  </picture>
131
111
 
132
- ---
133
-
134
- ## Why you can trust it
135
-
136
- Most scanners drown you in false positives until you stop reading them. Sniff is built the opposite way.
137
-
138
- We measured it on a fixture app planted with **21 bugs across all 12 classes**, plus a clean control page that should produce zero findings:
112
+ - **A terminal report** grouped by severity, each finding with steps, a fix, and a screenshot path.
113
+ - **A shareable file**, a self-contained HTML report with `--report` or JSON with `--json`.
114
+ - **An exit code**, non-zero when findings reach the `--fail-on` severity, so CI fails on real bugs.
115
+ - **A confidence label** on each. `uncertain` is hidden from the terminal unless you pass `--all`.
139
116
 
140
- | | Old engine | New engine |
141
- |:--|:--|:--|
142
- | Bugs found | 9 / 21 (43%) | **21 / 21 (100%)** |
143
- | Precision | ~13% | **100%** |
144
- | False positives | 125 | **0** |
145
- | Findings on the clean page | — | **0** |
146
- | Flagship command | crashed | works |
117
+ New in 0.8.0:
147
118
 
148
- Those numbers are locked as a regression test. The full suite is **441 tests**.
119
+ - `--caps scan,report` narrows the MCP server to the source scan and the saved-results reader, with
120
+ no browser launch and no browser download.
121
+ - `--storage-state auth.json` walks a logged-in app. Cookie and token values from that file are
122
+ redacted from the text of every written report, though not from screenshot pixels.
123
+ - An `assert` block in `sniff.config` caps findings by severity (`maxCritical`, `maxHigh`,
124
+ `maxTotal`), enforced on the command line by the walk, the source scan, and discovery.
149
125
 
150
- **How it keeps false positives near zero:**
151
-
152
- - A first-party **noise filter** drops the junk that isn't your bug — favicons, analytics, hot-module-reload chatter, expected auth redirects, engine aborts.
153
- - Accessibility findings are backed by **axe-core**, which is zero-false-positive by design.
154
- - **Uncertain findings are suppressed by default** (use `--all` to see them).
155
- - A broken page is reported **once**, not re-flagged on every link that points to it.
156
-
157
- If Sniff can't prove a bug, it doesn't claim one.
158
-
159
- ---
126
+ ## Works in your editor
160
127
 
161
- ## How is it different?
128
+ Works in Claude Code, Cursor, Codex, Copilot, Gemini CLI, and 70+ other agents through
129
+ `npx skills add`. The skills are Markdown, so they run on whatever model your editor points at.
162
130
 
163
- Linters read your source. End-to-end frameworks make *you* write the tests. Link checkers only check links. Sniff drives your real app and judges the result.
164
-
165
- | | **Sniff** | linkinator | pa11y | Playwright codegen | QA-Wolf-style services |
166
- |:--|:--|:--|:--|:--|:--|
167
- | Walks real user flows in a browser | **Yes** | No | No | You script it | Yes |
168
- | Zero setup, zero test-writing | **Yes** | Yes | Yes | No (you write tests) | No (onboarding) |
169
- | Broken links | **Yes** | Yes | No | Manual | Manual |
170
- | Accessibility (axe-core) | **Yes** | No | **Yes** | Manual | Some |
171
- | Empty / placeholder / fake data | **Yes** | No | No | No | No |
172
- | State-loss (back-button wipes a form) | **Yes** | No | No | Manual | Manual |
173
- | One-shot reproduction proof per finding | **Yes** | No | Partial | No | Varies |
174
- | Self-contained HTML report | **Yes** | No | Partial | No | Dashboard |
175
- | Runs locally, no account, no API key | **Yes** | Yes | Yes | Yes | No (service) |
176
-
177
- What Sniff uniquely does in one command: catch **empty/placeholder data**, **state-loss**, and **broken async outcomes** — and hand you a single proof report — with **no scripts to write and no service to sign up for**.
178
-
179
- ---
180
-
181
- ## Commands
182
-
183
- ```
184
- sniff Walk your app (auto-detects the dev server). The default.
185
- sniff --url <url> Walk a specific URL
186
- sniff scan Source-only scan, no browser (placeholders, TODOs, dead links, etc.)
187
- sniff report Show the results from the last run
188
- sniff doctor Check your environment (Node, browser, config, dev server)
189
- sniff ci Generate a GitHub Actions workflow
190
- sniff fix Auto-fix safe issues (console.log, debugger, etc.)
191
- sniff --help Show every command and flag
192
- sniff --version Show the version
193
- ```
194
-
195
- ### Useful flags
196
-
197
- | Flag | What it does |
198
- |:-----|:-------------|
199
- | `--url <url>` | Walk this URL instead of auto-detecting |
200
- | `--report` | Write a self-contained HTML report to `sniff-reports/sniff-report.html` |
201
- | `--all` | Also show low-confidence (`uncertain`) findings |
202
- | `--max-pages <n>` | Cap how many pages to walk (default: 25) |
203
- | `--no-mobile` | Skip the 375px responsive pass |
204
- | `--headed` | Show the browser window while it walks |
205
- | `--json` | Machine-readable JSON output |
206
- | `--ci` | CI mode (stable output, non-interactive) |
207
- | `--fail-on <sev>` | Exit non-zero on findings at or above this severity |
208
-
209
- ---
210
-
211
- ## Demo
212
-
213
- A real run against a buggy app — 21 real issues, zero false positives, every finding with severity, confidence, reproduction steps, and a fix:
214
-
215
- <img alt="Sniff walking a buggy app and reporting 21 real issues with zero false positives" src="https://raw.githubusercontent.com/Aboudjem/sniff/main/.github/assets/demo.svg" width="100%">
216
-
217
- ---
218
-
219
- ## Use it from your AI editor
220
-
221
- Sniff ships as an MCP server too. Add it once, then just ask your assistant *"scan this project for bugs"* or *"walk my app."* If your app is running, Sniff auto-detects it — you don't have to pass a URL.
222
-
223
- **One unified `sniff` tool, three modes:**
224
-
225
- - `walk` — **recommended.** Walks your running app's real flows (the flow-walk above).
226
- - `scan` — source-only scan, no browser.
227
- - `report` — show the last run's results.
228
-
229
- (`run` and `discover` are legacy modes kept for back-compat.)
131
+ | Agent | One-line install |
132
+ |:--|:--|
133
+ | Claude Code | `claude plugin install sniff@10x` |
134
+ | Any of 70+ agents | `npx skills add Aboudjem/sniff` |
135
+ | Codex, Gemini CLI, OpenCode, Pi | `./install.sh codex` |
136
+ | VS Code (Copilot) | `./install.sh copilot` |
137
+ | Everything else | see [docs/editors.md](docs/editors.md) |
230
138
 
231
139
  <details>
232
- <summary><b>Claude Code</b></summary>
233
-
234
- One-command plugin install from the [10x marketplace](https://github.com/Aboudjem/10x):
235
-
236
- ```bash
237
- claude plugin marketplace add Aboudjem/10x
238
- claude plugin install sniff@10x
239
- ```
240
-
241
- Or add just the MCP server:
140
+ <summary>Add it as an MCP server instead</summary>
242
141
 
243
142
  ```bash
244
143
  claude mcp add sniff-qa npx -- -y sniff-qa --mcp
245
- ```
246
- </details>
247
-
248
- <details>
249
- <summary><b>Cursor</b></summary>
250
-
251
- Add to `~/.cursor/mcp.json`:
252
-
253
- ```json
254
- { "mcpServers": { "sniff-qa": { "command": "npx", "args": ["-y", "sniff-qa", "--mcp"] } } }
255
- ```
256
- </details>
257
-
258
- <details>
259
- <summary><b>VS Code (Copilot)</b></summary>
260
-
261
- Add to `.vscode/mcp.json`:
262
-
263
- ```json
264
- { "servers": { "sniff-qa": { "type": "stdio", "command": "npx", "args": ["-y", "sniff-qa", "--mcp"] } } }
265
- ```
266
- </details>
267
-
268
- <details>
269
- <summary><b>Codex CLI</b></summary>
270
-
271
- ```bash
272
144
  codex mcp add sniff-qa -- npx -y sniff-qa --mcp
273
145
  ```
274
- </details>
275
146
 
276
- <details>
277
- <summary><b>Gemini CLI</b></summary>
278
-
279
- Add to `~/.gemini/mcp_config.json`:
280
-
281
- ```json
282
- { "mcpServers": { "sniff-qa": { "command": "npx", "args": ["-y", "sniff-qa", "--mcp"] } } }
283
- ```
147
+ Cursor, VS Code, Gemini CLI, Windsurf, Continue, OpenCode, and Zed take the same command as a JSON
148
+ or TOML entry. Every per-editor snippet is in [docs/editors.md](docs/editors.md).
284
149
  </details>
285
150
 
286
- <details>
287
- <summary><b>Windsurf</b></summary>
151
+ ## Good to know
288
152
 
289
- Add to `~/.codeium/windsurf/mcp_config.json`:
153
+ > [!IMPORTANT]
154
+ > No API key, no account, no signup, and no AI provider unless you set one up yourself. Walking and
155
+ > scanning never edit your source. `sniff fix` is the only command that rewrites code, and only when
156
+ > you run it.
290
157
 
291
- ```json
292
- { "mcpServers": { "sniff-qa": { "command": "npx", "args": ["-y", "sniff-qa", "--mcp"] } } }
293
- ```
294
- </details>
158
+ > [!NOTE]
159
+ > A walk clicks buttons and submits real forms, so it can create real data. Point it at a dev or
160
+ > staging app, not at production. The first walk also downloads a Chromium build and caches it, so
161
+ > that one run needs internet access.
295
162
 
296
- <details>
297
- <summary><b>Continue.dev</b></summary>
163
+ - **It wants a running app.** With no dev server up it falls back to a source-only scan and tells
164
+ you how to start the real walk. `npx sniff-qa scan` runs that scan on purpose.
165
+ - **Dead-link checking follows external links,** so a walk makes requests to the third-party URLs
166
+ your own pages already link to.
167
+ - **A walk that finds bugs exits 1** on purpose, so CI fails the build. That is not a crash.
168
+ `--fail-on none` turns off the severity gate, though an `assert` budget can still fail the run.
298
169
 
299
- Add to `.continue/mcpServers/sniff-qa.yaml`:
170
+ ## Learn more
300
171
 
301
- ```yaml
302
- mcpServers:
303
- sniff-qa: { command: npx, args: ["-y", "sniff-qa", "--mcp"], type: stdio }
304
- ```
305
- </details>
306
-
307
- > The first browser-based walk downloads Chromium (~165 MB). Over MCP, Sniff returns a structured `needsSetup` payload instead of blocking the editor on a long download — run the install it shows you, then ask again.
172
+ - [docs/editors.md](docs/editors.md), install and MCP snippets for every supported agent
173
+ - [docs/authenticated-walks.md](docs/authenticated-walks.md), walking a logged-in app with `--storage-state`
174
+ - [docs/assert-budgets.md](docs/assert-budgets.md), capping findings by severity in `sniff.config`
175
+ - [docs/comparison.md](docs/comparison.md), how sniff differs from linters, link checkers, and E2E frameworks
176
+ - [docs/faq.md](docs/faq.md), the questions this page does not answer
177
+ - [CHANGELOG.md](CHANGELOG.md) · [CONTRIBUTING.md](CONTRIBUTING.md) · [LICENSE](LICENSE)
308
178
 
309
179
  ---
310
180
 
311
- ## How it works
312
-
313
- 1. **Find the app.** Sniff auto-detects your running dev server (or you pass `--url`).
314
- 2. **Walk the flows.** It opens pages in a headless browser and interacts with them like a user — clicking, filling forms, following links — across desktop and a 375px mobile pass.
315
- 3. **Watch everything.** It records console errors, failed network requests, broken renders, missing feedback, and accessibility issues as it goes.
316
- 4. **Filter the noise.** The first-party noise filter and axe-core drop the false positives; uncertain findings are held back.
317
- 5. **Report with proof.** Each surviving finding gets a severity, a confidence, reproduction steps, a screenshot, and a suggested fix — in the terminal and an optional HTML report.
318
-
319
- <img alt="How sniff works: crawl, act, assert, prove, report" src="https://raw.githubusercontent.com/Aboudjem/sniff/main/.github/assets/how-it-works.svg" width="100%">
320
-
321
- ---
322
-
323
- ## CI integration
324
-
325
- Run Sniff in your pipeline and fail the build on real bugs:
326
-
327
- ```bash
328
- npx sniff-qa --ci --fail-on high
329
- ```
330
-
331
- Generate a ready-to-commit GitHub Actions workflow:
332
-
333
- ```bash
334
- npx sniff-qa ci
335
- ```
336
-
337
- This writes `.github/workflows/sniff.yml` with browser caching and report artifacts.
338
-
339
- ---
340
-
341
- ## FAQ
342
-
343
- **Does it work without a dev server?**
344
- Sniff is built to walk a *running* app, so that's where it shines. If no server is running, it doesn't fail silently — it runs a source-only scan and tells you exactly how to start the real walk (start your dev server or pass `--url`). You can also run `npx sniff-qa scan` to get the source scan on purpose.
345
-
346
- **What gets downloaded on first run?**
347
- The first time Sniff opens a browser, it downloads a Chromium build (~165 MB, one-time, then cached). You'll see the progress, and you need internet access for that first run. Nothing else is installed and no account is created.
348
-
349
- **Do I need an API key?**
350
- No. Sniff runs entirely on your machine with no API key and no signup. Your code and your app never leave your computer.
351
-
352
- **How is it different from a linter?**
353
- A linter reads your source files and never runs your app, so it can't see a dead submit button, an infinite spinner, a wiped form, or a 500 page. Sniff opens your real app, interacts with it, and reports what actually broke — with a screenshot and steps to reproduce.
354
-
355
- **How is it different from Playwright codegen (or writing E2E tests)?**
356
- Playwright codegen records a script that *you* author and maintain; it tests only the path you clicked. Sniff writes nothing for you to maintain — it explores your flows on its own and judges the outcome, catching things a recorded happy-path never checks (empty/placeholder data, state-loss, missing success feedback).
357
-
358
- **Will it change my code?**
359
- No, not during a walk. Walking and scanning are read-only. The separate `sniff fix` command applies safe auto-fixes (like stray `console.log`/`debugger`) and only when you run it.
360
-
361
- **What stacks does it work with?**
362
- Any web app you can open in a browser — React, Next.js, Vue, Svelte, Angular, Remix, SvelteKit, Astro, plain HTML, and more. It walks the rendered app, so the framework doesn't matter for the browser checks.
363
-
364
- ---
365
-
366
- ## Works first-class in
367
-
368
- Claude Code · Cursor · VS Code (Copilot) · Codex · Gemini CLI · Windsurf · Continue.dev — via the MCP server (command `npx`, args `["-y", "sniff-qa", "--mcp"]`) or the CLI directly.
369
-
370
- ---
371
-
372
- ## Contributing
373
-
374
- Issues and PRs welcome. See [CONTRIBUTING.md](CONTRIBUTING.md).
375
-
376
- ---
377
-
378
- <p align="center">
379
- <sub>
380
- Built on <a href="https://playwright.dev">Playwright</a> · <a href="https://github.com/dequelabs/axe-core">axe-core</a> · <a href="https://developer.chrome.com/docs/lighthouse">Lighthouse</a> · <a href="https://github.com/mapbox/pixelmatch">pixelmatch</a> · <a href="https://zod.dev">Zod</a> · <a href="https://github.com/modelcontextprotocol/typescript-sdk">MCP SDK</a>
381
- </sub>
382
- </p>
383
-
384
- <p align="center">
385
- <a href="https://www.linkedin.com/in/adam-boudjemaa/"><img src="https://img.shields.io/badge/LinkedIn-0A66C2?style=flat-square&logo=linkedin&logoColor=white" alt="LinkedIn"></a>
386
- <a href="https://x.com/AdamBoudj"><img src="https://img.shields.io/badge/X-000000?style=flat-square&logo=x&logoColor=white" alt="X"></a>
387
- <a href="https://adam-boudjemaa.com/"><img src="https://img.shields.io/badge/Website-ef4444?style=flat-square&logo=googlechrome&logoColor=white" alt="Website"></a>
388
- </p>
389
-
390
- <p align="center">
391
- <sub>Built by <a href="https://github.com/Aboudjem">Adam Boudjemaa</a> · <a href="LICENSE">Apache 2.0</a></sub>
392
- </p>
181
+ <p align="center"><sub>Built by <a href="https://github.com/Aboudjem">Adam Boudjemaa</a> · <a href="LICENSE">Apache 2.0</a> · standing on <a href="https://playwright.dev">Playwright</a> and <a href="https://github.com/dequelabs/axe-core">axe-core</a></sub></p>