{"id":"18397df3-8191-4f8c-8e50-dd41f0353290","slug":"consentflow-debugger","kind":"mcp","name":"ConsentFlow-Debugger","summary":"AI-agent ready (MCP): ask Claude or Cursor to check any live page for tags that fire before consent, even after clicking accept or reject, and get the fix. Also a Chrome extension and a CLI. Read-only","owner":"ChristopherLandaverde","repoUrl":"https://github.com/ChristopherLandaverde/ConsentFlow-Debugger","license":"MIT","licenseTier":"permissive","trust":"community","stars":2,"match":null,"fewStars":true,"format":"mcp","createdAt":"2025-05-15T13:03:54.000Z","addedAt":"2026-11-09T14:17:40.553Z","logoUrl":"https://avatars.githubusercontent.com/u/22153509?u=70d74006d8cd2cd784639990236b6b6b80bbed3c&v=4","linkOnly":false,"collection":null,"skillPath":null,"categories":["browser-web","data-analytics"],"tags":["analytics","browser-extension","chrome-extension","consent-mode-v2","cookie-consent","developer-tools","google-tag-manager","gtm"],"runtime":"binary","local":true,"cloudCapable":false,"cloudCapability":"unknown","cloudReadyOnPlatform":false,"needsApiKey":false,"reviewed":false,"syncedAt":"2026-11-09T14:27:08.809Z","sample":false,"url":"https://conduits.dev/tools/consentflow-debugger","language":"JavaScript","transports":["stdio"],"install":{},"requirements":{"env":[],"upstreamMayCharge":false},"tools":null,"readme":"

\n \"ConsentFlow\n

\n\n

ConsentFlow Debugger

\n\n

\n Tags fire before consent on live pages. Find them and get the fix from the extension, the CLI or your AI agent.\n

\n\n

\n 8 Consent Mode rules · 10 tags judged request by request · MCP server for Claude, Cursor and any MCP client · the extension sends nothing anywhere\n

\n\n

\n \"CI\"\n \"agentic:\n \"Version\"\n \"License:\n

\n\n

\n Quick start · For agents · What you get · Known limits · Privacy · Landing page\n

\n\nConsentFlow Debugger is a free, open-source Chrome extension for analytics engineers, QA\ntesters and consultants. It checks Google Consent Mode on the live page in one click and gives\nthe fix for each problem. Your AI agent can run the same checks through a local MCP server:\ncheck a URL, explain a finding, list the rules. Every tool is read-only; fixing the tags or\nthe CMP stays with you.\n\nA Consent Mode setup can look right in GTM preview and still be wrong on the live page: the\nconsent default lands a few hundred milliseconds after `config`, a pixel fires before the\nbanner is answered, or the v2 fields `ad_user_data` and `ad_personalization` are missing.\nGoogle's own guide is blunt about the first one: \"If your consent code is called out of\norder, consent defaults won't work\"\n([source](https://developers.google.com/tag-platform/security/guides/consent)). The cost is\nrequests sent for visitors who said no, and consent signals that Google's\n[conversion modelling](https://support.google.com/google-ads/answer/10000067) relies on\nbeing wrong. ConsentFlow shows you this on the page itself, with the timings and requests as\nevidence.\n\n

\n \"Four\n
A Consent Mode bug caught, fixed and confirmed, on the two sample pages bundled in tests/fixtures/. The fixed sample also differs in more than the fix (a OneTrust stub, a consent update, no Meta Pixel), so frame 3 is the same checks on a correctly wired page.\n

\n\n## Agentic\n\n

\n \"ConsentFlow\n

\n\nThe extension's engine also runs headless, so an agent can check a page the way you would with\nthe popup. The agent picks the tool and explains the result; the verdict and every number come\nfrom `popup/findings.js`.\n\n| Ask your agent… | Tool | Reads | Writes |\n|---|---|---|---|\n| \"Check `http://localhost:8000/popup-states.html?s=broken` and explain the first problem.\" | `check_url` | The page, loaded in its own headless Chromium: consent commands and requests to known ad and analytics vendors | Nothing itself; the page makes its own requests as for any visitor |\n| \"Click Reject all on the banner at staging.example.com. Does anything still fire?\" | `check_url` with `consent: \"deny\"` | The same, after clicking the reject-all button of OneTrust, Cookiebot, Google Funding Choices or Usercentrics | Nothing on your site itself; the click happens in that throwaway browser, and the page and CMP send their own requests as for any visitor |\n| \"Why does `consent-default-before-config` matter, and where is it judged?\" | `explain_finding` | The finding's copy and the `popup/findings.js` lines that build it | Nothing |\n| \"Which tags can send cookieless pings without counting as a violation?\" | `list_checks` | The 8 rules and 11 tags in `injected.js` (only GA4, Google Ads and Floodlight can get `review`) | Nothing |\n\nAgentic levels: L1 machine-readable · L2 agent-callable · L3 live read. ConsentFlow: L1 (JSON report with `schemaVersion` and exit codes) · L2 (MCP\nserver) · L3 (fixture-verified: tested on the bundled sample pages and stand-in CMP banners, not\nyet on a public site or a vendor's real banner). It has no L4: nothing writes.\n\nPlanned: an L3 scenario matrix (several pages and consent states in one run) and a recorded run\non a real public site, which drops the fixture-verified label.\n\nDetails: [Use it from an agent / CLI](#use-it-from-an-agent--cli) (the JSON report, options and\nexit codes) · [MCP server](#mcp-server-ask-your-agent) (setup for Claude Code, Claude Desktop\nand Cursor, errors, and what goes where).\n\n## Who it's for\n\n- **Analytics engineers** verifying a Consent Mode change on staging or production before they call it done.\n- **QA testers** reproducing a consent bug and filing it with the timings and requests as evidence.\n- **Consultants** auditing a client's site from the browser, without access to their GTM container.\n\n## What you get\n\n| | |\n|---|---|\n| **A verdict in one sentence** | `× BROKEN`, `! REVIEW`, `– INCOMPLETE` or `✓ VERIFIED`, with the most severe problem underneath. Checks that couldn't run say so and why; they never count as passed. |\n| **Every tag judged by the consent at the moment it sent** | A tag that sent a request while its consent was `denied` is a violation, even if the visitor accepted later. Google's cookieless pings under advanced Consent Mode are flagged for review instead. Each request is judged only against the tag that sent it, read from its tag ID or endpoint, so a GA4 hit is never blamed on Google Ads. Covers GA4, Google Ads, Floodlight, Meta Pixel, LinkedIn, TikTok, Hotjar, Clarity, Pinterest and Snapchat, and the tags inside a GTM container. |\n| **Ordering and setup mistakes, timed in milliseconds** | `config` before the consent default, an `update` before the default, a missing default, duplicate defaults, missing v2 fields, no `wait_for_update` when a CMP is present. |\n| **A fix and a way to check it** | Every finding says what was observed, why it matters, how to fix it (page source and GTM) and what to look for after you reload. |\n| **Google's consent signals decoded** | `gcs` and `gcd` on each Google request, side by side with the consent the page had when it was sent, flagged when they disagree. |\n\nThe full detection reference, including all eight rules and the tag verdicts, is in\n[docs/detection.md](docs/detection.md).\n\n## Quick start\n\nNot published on the Chrome Web Store: you load it unpacked, with no build step. Requires\nChrome or another Chromium browser with Manifest V3 (Chrome 114+ for the side panel), and\nPython 3.7+ to serve the sample pages.\n\n```bash\ngit clone https://github.com/ChristopherLandaverde/ConsentFlow-Debugger.git\ncd ConsentFlow-Debugger\npython3 -m http.server 8000 -d tests/fixtures\n```\n\n1. In `chrome://extensions/`, turn on **Developer mode**, click **Load unpacked** and select\n the `ConsentFlow-Debugger` folder (the one with `manifest.json`).\n2. Open and click the ConsentFlow icon.\n You should see `× BROKEN`, \"2 consent problems on this page\", starting with \"Google tag\n configured before the consent default\".\n3. Open : the popup says \"Consent is\n wired correctly on this page.\"\n4. Now reload a page of your own and click the icon. Click **Open full trace** for the\n timeline, every request and the consent state in a side panel.\n\nContent scripts don't run on `file://` pages, which is why the samples are served over\nHTTP. No local server? The landing page's\n[measurement test section](https://consentflow-psi.vercel.app/#measurement-playground) loads\na disposable GTM container. To update, `git pull` and click reload on the extension's card.\nRelease-zip install and a full tour of the popup and side panel, with screenshots:\n[docs/usage.md](docs/usage.md).\n\n## Use it from an agent / CLI\n\n`scripts/consentflow-check.mjs` checks a URL in headless Chromium with the extension's own\ncode and prints the verdict as JSON, so a script, CI job or coding agent can read it. It adds\n`injected.js` to the page before any page script runs (top frame only, like the extension),\nreads the state the way `content.js` does, and builds the verdict and findings with\n`popup/findings.js`, the file the popup renders. Verified on the bundled fixtures served\nlocally; not yet run against a public site.\n\n```bash\nnpm ci --ignore-scripts\nnpx playwright install chromium # once: the browser the runner drives\npython3 -m http.server 8000 -d tests/fixtures &\nnode scripts/consentflow-check.mjs \"http://localhost:8000/popup-states.html?s=broken\" \\\n | jq '{schemaVersion, tool, consentMode, verdict: .verdict.label, title: .verdict.title, first: .findings[0], requests: (.requests | length)}'\n```\n\nOutput from that run (exit code 1). The `+ms` timings are measured on each load, so they\nvary from run to run:\n\n```json\n{\n \"schemaVersion\": \"1\",\n \"tool\": \"consentflow\",\n \"consentMode\": \"none\",\n \"verdict\": \"× BROKEN\",\n \"title\": \"2 consent problems on this page.\",\n \"first\": {\n \"id\": \"consent-default-before-config\",\n \"status\": \"broken\",\n \"title\": \"Google tag configured before the consent default\",\n \"observed\": \"`config` for G-FIXTURE ran at +73 ms; the consent default ran at +398 ms.\",\n \"consequence\": \"Requests in that 325 ms gap assume consent is granted.\",\n \"fix\": [\n \"Move the consent default above the Google tag snippet in the page source.\",\n \"In GTM, fire the consent default on the Consent Initialization – All Pages trigger.\"\n ],\n \"verify\": \"Reload. The consent default should come before `config` in the timeline.\",\n \"evidence\": [\n { \"label\": \"First config\", \"value\": \"G-FIXTURE at +73 ms\" },\n { \"label\": \"Consent default\", \"value\": \"+398 ms\" },\n { \"label\": \"Gap\", \"value\": \"325 ms\" }\n ]\n },\n \"requests\": 2\n}\n```\n\nThe full report is `{ schemaVersion, tool, url, finalUrl, capture, consentMode,\nconsentInteraction, consentButton, navigatedAfterConsent, verdict, findings, requests }`.\n`capture` is `complete` when ConsentFlow finished its first pass on the page; if it didn't\n(`not_ready`, or `no_response` when the page stopped answering), `verdict` is `null` and\n`findings` is empty, because there is nothing to judge. Each finding has `id`, `status`, `title`,\n`observed`, `consequence`, `fix`, `verify` and `evidence`; each request has when it was sent,\nits vendor, host, path, the tag it is attributed to, `gcs` / `gcd`, and the consent on the\npage at that moment. `--format text` prints the same verdict as lines instead.\n\n| Option | |\n|---|---|\n| `--consent none` (default) | Load the page and touch nothing. This is the mode that catches tags firing before consent. |\n| `--consent accept` / `deny` | Click the accept-all or reject-all button of OneTrust, Cookiebot, Google Funding Choices or Usercentrics, using known button selectors. Each is tested only on a stand-in banner (`tests/fixtures/cmp-banner.html`), not on the vendor's real one. With none of those buttons on the page it reports `\"consentInteraction\": \"not_found\"` and clicks nothing; a button it can't click (covered by another element, say) is `\"click_failed\"`. If a new document loads during the wait after the click (a navigation or reload), `navigatedAfterConsent` is `true` and the capture covers that new page load. A URL change without a new document (`pushState`, a hash change) leaves it `false`. |\n| `--wait ` | How long to keep capturing after the page loads, and again after a click. Default 3000. |\n\nExit codes: `0` VERIFIED or REVIEW; `1` BROKEN; `3` INCOMPLETE, a capture that didn't\nfinish, or a requested banner click that didn't happen (`not_found` or `click_failed`, even\nwhen the page is also BROKEN, since the scenario you asked for wasn't run); `2` bad arguments,\na page that couldn't be loaded or answered HTTP 400 or above, or the runner itself failing\n(for example, Chromium not installed).\n\nRun time: a normal page takes a few seconds plus `--wait` (twice `--wait` after a click).\nEvery step has a time limit, so with the defaults a run that hits all of them ends within about\n57 s for `--consent none` and 71 s for `accept` / `deny`, plus browser start-up: 31 s to load\nthe page, 11 s for ConsentFlow's first pass, the 3 s wait, 6 s to read the state and 6 s to close\nthe browser, and for a click up to 11 s more for the banner and another 3 s wait.\n\nWhat it doesn't do: it checks one page load from wherever you run it, so visitor regions are\nnot simulated (a CMP that shows its banner only in the EU may not show it to a run from\nelsewhere); only the four CMPs above can be clicked, and their selectors have not been checked\nagainst the real banners; capture starts slightly earlier than in the extension, whose\nscript loads asynchronously; and the [known limits](#known-limits) below apply as they do in\nthe extension. The runner sends nothing anywhere itself: the browser loads the page you\nname, and that page makes its own requests as it would for any visitor.\n\n### MCP server: ask your agent\n\n`scripts/consentflow-mcp.mjs` is a local stdio [MCP](https://modelcontextprotocol.io) server\nover the same runner, so Claude Code, Claude Desktop, Cursor or any MCP client can check a page\nand explain what it found. It runs from source on your machine; it is not published to npm or\nan MCP registry. All three tools are read-only:\n\n| Tool | Input | Returns |\n|---|---|---|\n| `check_url` | `url` (http or https only), `consent?` `none` / `accept` / `deny`, `waitMs?` (0 to 60000) | The runner's JSON report above, unchanged. |\n| `explain_finding` | `id`, e.g. `consent-default-before-config` or `tag:Facebook Pixel` | The finding's statuses, the observed / consequence / fix / verify / evidence copy `popup/findings.js` gives where it quotes no measured value, and the `findings.js` source (file, lines, code) that builds the rest. |\n| `list_checks` | none | The 8 rules and 11 tags the engine knows, read from `injected.js`, each with the statuses it can give and its source location. Only GA4, Google Ads and Floodlight can get `review` (a cookieless ping under advanced Consent Mode); for any other tag, a request sent while its consent was denied is `broken`. |\n\nThe verdict and every number come from `popup/findings.js` through the runner; the server only\nchecks input and shapes errors. Errors come back as tool errors with\n`{ \"error\": { \"code\", \"message\" } }`: `invalid_url` (anything but http/https, such as `file:`,\n`data:` or `chrome:`), `navigation_failed`, `capture_not_ready` / `capture_no_response`\n(with the report, which has no verdict), `runner_failed` (for example, Chromium not installed),\n`busy` (more than 2 `check_url` runs at once; each starts its own Chromium) and\n`unknown_finding`. Input that doesn't match a tool's schema (an unknown `consent` value, a\nnegative `waitMs`, no `url`) is refused by the MCP SDK before the tool runs: a tool error whose\ntext starts `MCP error -32602: Input validation error`. A requested banner click that didn't\nhappen is not an error: the report comes back with `consentInteraction` `not_found` or\n`click_failed`. Stdout carries JSON-RPC only; every log goes to stderr.\n\n**What goes where.** The server sends nothing anywhere itself, and Chromium loads the page you\nname as the CLI does. What the tools return (the report, with the page's URL, findings and\nrequest hosts and paths) goes to your MCP client, and through it to the model behind your\nagent and that model's provider. Don't check pages whose URLs or requests you can't share with\nthat provider. Three more things to know:\n\n- `check_url` loads any http(s) URL from your machine, including `localhost` and private\n network addresses. That is how you check a local or staging build, but it also means an agent\n following instructions planted in content it read (a prompt injection) could make your\n machine load an internal page. What comes back is limited to the report: `finalUrl`, the\n findings, and the host and path of each request to a known ad or analytics vendor, not page\n content.\n- Strings taken from the page (tag and config IDs, CMP names, request hosts and paths, the final\n URL) reach the model as they are, unfiltered.\n- The URL is echoed back as given (normalized), so credentials or tokens in it (`user:pass@`,\n a `?token=` query) go to the model too. Don't pass URLs that carry secrets.\n\nSetup, after `npm ci --ignore-scripts` and `npx playwright install chromium` in this repo.\nUse absolute paths. If you use nvm, `node` may not be on the PATH the client starts servers\nwith, so give the absolute path from `which node`.\n\nClaude Code (all projects):\n\n```bash\nclaude mcp add -s user consentflow -- \"$(which node)\" \"$PWD/scripts/consentflow-mcp.mjs\"\nclaude mcp list # consentflow: … - Connected\n\n# If Playwright's browsers are not in the default cache (see below):\nclaude mcp add -s user consentflow -e PLAYWRIGHT_BROWSERS_PATH=/path/to/browsers -- \"$(which node)\" \"$PWD/scripts/consentflow-mcp.mjs\"\n```\n\nClaude Desktop (`~/Library/Application Support/Claude/claude_desktop_config.json` on macOS,\n`%APPDATA%\\Claude\\claude_desktop_config.json` on Windows) and Cursor (`~/.cursor/mcp.json`)\ntake the same block:\n\n```json\n{\n \"mcpServers\": {\n \"consentflow\": {\n \"command\": \"/Users/you/.nvm/versions/node/v22.12.0/bin/node\",\n \"args\": [\"/Users/you/ConsentFlow-Debugger/scripts/consentflow-mcp.mjs\"]\n }\n }\n}\n```\n\nMCP clients start servers with a reduced environment. If Playwright's browsers are somewhere\nother than the default cache (`PLAYWRIGHT_BROWSERS_PATH` is set in your shell), add\n`\"env\": { \"PLAYWRIGHT_BROWSERS_PATH\": \"/that/path\" }` to the block, or use the second\n`claude mcp add` command above (`-e` goes after the server name); without it `check_url` returns\n`runner_failed` with Playwright's \"Executable doesn't exist\" message.\n\nThen ask, for example: \"Use consentflow to check http://localhost:8000/popup-states.html?s=broken\nand explain the first problem.\"\n\n
\nTool calls from a real session on the bundled broken fixture (local, python3 -m http.server 8000 -d tests/fixtures), trimmed\n\nRecorded while building the server: an MCP client (the SDK's `Client` over stdio) started\n`node scripts/consentflow-mcp.mjs` and made these calls. Results are cut down to the fields\nshown; the `+ms` timings vary per load.\n\n```text\n→ tools/list\n← check_url, explain_finding, list_checks\n\n→ tools/call check_url {\"url\": \"http://localhost:8000/popup-states.html?s=broken\"}\n← isError: false\n verdict: {\"status\": \"broken\", \"label\": \"× BROKEN\", \"title\": \"2 consent problems on this page.\",\n \"detail\": \"Also 4 checks not run yet.\",\n \"counts\": {\"broken\": 2, \"review\": 0, \"notChecked\": 4, \"verified\": 4, \"total\": 10}}\n findings[].id/status: consent-default-before-config broken · tag:Facebook Pixel broken ·\n consent-update-received, gcs-consistency, default-before-update, wait-for-update not_checked ·\n consent-default-exists, v2-types-present, no-duplicate-defaults, tag:Google Analytics 4 verified\n findings[0]: \"observed\": \"`config` for G-FIXTURE ran at +61 ms; the consent default ran at +391 ms.\"\n \"consequence\": \"Requests in that 330 ms gap assume consent is granted.\"\n \"evidence\": First config = G-FIXTURE at +61 ms · Consent default = +391 ms · Gap = 330 ms\n requests: 2\n\n→ tools/call explain_finding {\"id\": \"tag:Facebook Pixel\"}\n← statuses: broken, verified, not_checked\n verdicts: violation, blocked, firing, idle\n copy.blocked: {\"status\": \"verified\", \"title\": \"Facebook Pixel held back while ad_storage is denied\",\n \"observed\": \"`ad_storage` is denied, and no requests to Meta went out.\", …}\n copySource: popup/findings.js lines 431-552, covers violation, firing\n\n→ tools/call check_url {\"url\": \"file:///etc/passwd\"}\n← isError: true\n {\"error\": {\"code\": \"invalid_url\", \"message\": \"Only http and https URLs can be checked.\", \"url\": \"file:///etc/passwd\"}}\n\n→ tools/call list_checks {}\n← 8 rules, 11 tags; cookielessPings true for Google Analytics 4, Google Ads, Floodlight\n```\n\n
\n\n## Known limits\n\n- Requests are captured when made through `fetch`, `XMLHttpRequest` or\n `navigator.sendBeacon`. Image and iframe requests are seen only for Google hosts (Floodlight\n activity tags, some Google Ads hits), through the page's Resource Timing entries, which\n appear once the load finishes; one that never finishes may be missed. Other vendors' image\n pixels and all script loads are not seen, so a vendor that only uses `` beacons can show\n as `IDLE` or `BLOCKED` while it is in fact sending data.\n- Google's cookieless pings under advanced Consent Mode are a `! REVIEW` finding, not a\n violation, when the request's own `gcs` / `gcd` reports its storage type as denied and it\n carries no advertising cookie ID (`auid`, `auiddc`, `gclaw`). ConsentFlow can't see request\n headers or read cookies, so for GA4 it trusts the request's own `gcs`. The rule is in\n [docs/detection.md](docs/detection.md#cookieless-pings-advanced-consent-mode).\n- A Google request with no tag ID on an endpoint several Google tags share (such as\n `ad.doubleclick.net/ccm/s/collect`) is listed as \"Not tied to a tag\" in the side panel's\n Requests view and changes no verdict. See\n [docs/detection.md](docs/detection.md#which-tag-sent-a-request).\n- Only the top-level frame is inspected; iframes are not.\n- State is kept per page load. Reload the page to start a fresh capture.\n- If a tab was already open when the extension was installed or reloaded, the content script\n is injected on demand but misses everything that happened earlier. Reload the tab.\n- The timeline keeps the latest 300 events and the latest 200 network requests.\n\n## Privacy and permissions\n\nNothing leaves your browser. The extension's own code makes no network requests: it has no\nanalytics, no remote configuration and no server. It stores no captured data: that lives only\nin the inspected page's memory and the open popup or side panel, and is gone when the page is\nclosed or reloaded. The only thing it stores is UI state (what you expanded, which side-panel\ntab you picked) in `chrome.storage.session`, which Chrome keeps in memory and clears when the\nbrowser closes.\n\nThis is about the extension. The [headless runner and MCP server](#use-it-from-an-agent--cli)\nrun outside the browser; through MCP, what they report reaches the model behind your agent\n([what goes where](#mcp-server-ask-your-agent)).\n\nTwo things to know:\n\n- The capture script runs in the page's own JavaScript context (it has to, to see\n `dataLayer` and `gtag`), so scripts on the inspected page can read what the extension\n captured about that page (`window.__cfDebug`).\n- It wraps `fetch`, `XMLHttpRequest.prototype.open`, `navigator.sendBeacon` and\n `dataLayer.push` to observe them, and defines `gtag` if the page has none. Calls are passed\n through unchanged. It also reads the page's Resource Timing entries (a `PerformanceObserver`)\n for the URLs of image and iframe loads to Google hosts.\n\n| Permission | Why |\n|---|---|\n| Content script on all `http`/`https` pages at `document_start` | Consent defaults are often set before GTM loads, so capture must start first. Chrome warns it can \"read and change all your data on all websites\" because of this. |\n| `activeTab`, `scripting` | Reach the current tab, and inject into tabs opened before the extension was installed. |\n| `sidePanel` | The **Open full trace** side panel. |\n| `storage` | `chrome.storage.session` for UI state only. |\n\nWhen something is broken or needs review, a \"Discuss a system review ↗\" link appears below\nall findings; it opens `https://focosys.io/review` and sends no page data. Full detail:\n[docs/privacy.md](docs/privacy.md) and [SECURITY.md](SECURITY.md).\n\n## How it works\n\n`injected.js` runs in the page, captures consent commands and vendor requests, and runs the\nrules; `content.js` relays its state to the popup and side panel; `background.js` sets the\ntoolbar badge. File roles and a sequence diagram: [docs/architecture.md](docs/architecture.md).\nThis repo also holds the landing page ([docs/website.md](docs/website.md)) and some legacy\nCookiebot helper scripts that are not part of the extension.\n\n## Development\n\nRequires Node.js 22. The extension has no dependencies. The dev dependencies are Playwright,\nwhich runs the headless runner's tests in Chromium, and the MCP SDK and zod for the MCP server.\n\n```bash\nnpm ci --ignore-scripts\nnpx playwright install chromium # once, for the runner's tests\nnpm test # runs the real injected.js and popup code in node:vm\nnpm run validate # required files, manifest content script, node --check\n```\n\nWhat the tests cover and what they don't: [docs/development.md](docs/development.md).\n\n## Contributing\n\nIssues and pull requests are welcome. Branch from `main`, keep `content.js` talking to\n`injected.js` only through `window.postMessage`, run the three commands above, and say in the\nPR which pages you checked with the extension loaded. If you change what a rule or verdict\nmeans, explain the Consent Mode behaviour it is based on. More in\n[docs/development.md](docs/development.md#contributing-in-detail).\n\n## License\n\nMIT. See [LICENSE](LICENSE).\n\nBuilt by [Christopher Landaverde](https://focosys.io).\n"}