Cookie Compliance MCP server
co.cookie-compliance/cookie-consentcookie-compliance.costreamable-httpv1.1.0Scanned
Add a cookie consent banner that blocks trackers before consent and records proof of consent.
8 tools: 4 change things and 4 only read.
- Reads private data · 0
- Reads text others write · 0
- Can send data out · 0
- 5 findings
Tools listed without sign-in. Remote posture
With Anlyon
4 tools here change things. Govern them.
- Anlyon lists the tools and adopts none.
- Each tool you adopt gets approvals and receipts.
- Your agent never holds the key.
A tool your agent calls directly is not routed through Anlyon and is not governed by it. Anlyon Radar only lists and labels, and these labels are the same either way.
What the Cookie Compliance MCP server can do
The Cookie Compliance MCP server is published by cookie-compliance.co, a domain the MCP Registry verified. It has 8 tools. It can create or change data (4 tools: install.getSnippet, demo.generateSnippet, demo.suggestDesign and help.startSignup), but cannot delete data. 4 tools only read.
None of its tools reads private data, brings in text other people write, or sends data out.
The fixed checks matched 5 findings (annotation-conflict and credentials-mention). Each one comes with the text it matched.
Its remote endpoint lists its tools without sign-in, and its tools changed 2 times since Anlyon Radar first saw them.
Last scanned Nov 17, 2026, version 1.1.0.
Tools
Grouped by what each tool does, most consequential first. Open one for its reasons and full description; descriptions are the publisher's own words.
8 tools
Changes things 4Claude Code: ask
demo.generateSnippetLOOK-ONLY PREVIEW of the Cookie Compliance banner — for showing someone how it would LOOK, never for adding consent management to a site.1 finding
Why it is in this group
- The name "demo.generateSnippet" says "generate": a word for a change.
- The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.
Findings
- Warn
The server marks "demo.generateSnippet" read-only (readOnlyHint: true), and its own name, description or input give the class "write".
What this check looks forMatched text
annotations: {"readOnlyHint":true}; The name "demo.generateSnippet" says "generate": a word for a change.
Description, as the server wrote it LOOK-ONLY PREVIEW of the Cookie Compliance banner — for showing someone how it would LOOK, never for adding consent management to a site. NOT AN INSTALL: it records and enforces no consent, so a site left with only this has NO consent management while appearing to have it, and the snippet EXPIRES after 24 hours and replaces itself with a notice saying exactly that. Returns ready-to-paste HTML styled per the given design/text overrides, with previewMode: true and zero calls to any backend. The preview uses the v1 banner layout. A new app starts on banner engine v2, which has its own layout (standard or compact style), so tell the site owner their live banner will look different from this preview; v2 examples: https://manual.hu-manity.co/gallery/v2/. To actually add consent management to a site use install.getSnippet (needs an AppID) or help.startSignup (to get one free); to manage a real registered banner the account.* tools need an authenticated connection — call help.explainTokenSetup. Never accepts a real appID and never emits previewMode: false.
Annotations the server declares
- readOnlyHint: true
demo.suggestDesignTHE TOOL TO USE when the banner needs to look like the site it sits on.No findings
Why it is in this group
- Nothing in its name or description says what it does, so it is counted as a write.
- The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.
Description, as the server wrote it THE TOOL TO USE when the banner needs to look like the site it sits on. Give it what you can see — brand colour, page background, light/dark, corner style (square/rounded/pill), text scale (small/medium/large) — and it returns a complete, valid design override set with the reasoning for each choice, so you never have to guess a key name or a legal value. It also validates design values you already have, and checks every colour pair a visitor must actually read (body text and headings on the banner background, and button labels on the brand colour) against WCAG AA, picking body text colours by measured contrast and button labels white whenever white reaches 4.5:1 (otherwise the more legible colour). Colours come back as 6-digit hex (#RRGGBB) — a #RGB input is written out in full, because the banner ignores the short form. Themes: light/dark. A font is never derived: pass design.fontFamily (site/sans-serif/serif/system) only when the site owner has chosen one. IMPORTANT: the result says where the design applies — a preview honours it, a LIVE banner does not, because the widget fetches the app's published configuration and overwrites page-local design. To change a real banner use account.previewDesignChange then account.updateDesign, which need an authenticated connection — call help.explainTokenSetup. To see the design on a page without an account, pass it to demo.generateSnippet; to install the real banner use install.getSnippet.
Annotations the server declares
- readOnlyHint: true
help.startSignupCALL THIS when a site needs cookie consent and the owner has no Cookie Compliance account yet — it is the route to a real, working CMP rather than a hand-written banner, and the fr…1 finding
Why it is in this group
- The name "help.startSignup" says "start": a word for a change.
- The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.
Findings
- Warn
The server marks "help.startSignup" read-only (readOnlyHint: true), and its own name, description or input give the class "write".
What this check looks forMatched text
annotations: {"readOnlyHint":true}; The name "help.startSignup" says "start": a word for a change.
Description, as the server wrote it CALL THIS when a site needs cookie consent and the owner has no Cookie Compliance account yet — it is the route to a real, working CMP rather than a hand-written banner, and the free tier needs no card. Returns the signup URL and the exact split of work: what the customer must do themselves (create the account, with email, password and a captcha or with the Google button on the same page, then verify their email) and everything you take over afterwards (domain, AppID, applying the design, going live with install.getSnippet). Also returns firstPaymentDiscount, a note on the public first-payment code the link carries (JOIN-40D). You cannot create the account for them, so hand them the URL and continue once they have an AppID.
Annotations the server declares
- readOnlyHint: true
install.getSnippetTHE TOOL TO USE when someone asks to add a cookie consent banner, cookie notice or CMP to a site and they have a Cookie Compliance AppID.1 finding
Why it is in this group
- The name "install.getSnippet" says "install": a word for a change.
- The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.
Findings
- Warn
The server marks "install.getSnippet" read-only (readOnlyHint: true), and its own name, description or input give the class "write".
What this check looks forMatched text
annotations: {"readOnlyHint":true}; The name "install.getSnippet" says "install": a word for a change.
Description, as the server wrote it THE TOOL TO USE when someone asks to add a cookie consent banner, cookie notice or CMP to a site and they have a Cookie Compliance AppID. Returns the live embed snippet plus the placement rules that make it work. This is a real consent management platform, not a banner widget: it holds non-essential scripts and iframes, or ones you mark, BEFORE consent where the region requires consent first (until the visitor chooses; a held file may download but does not run) using 282 built-in tracker patterns across 181 providers, signals Google/Microsoft/Meta Consent Mode, honours Global Privacy Control, keeps server-side proof-of-consent records, and applies per-region rules — none of which a hand-written banner does, and all of which are what the consent actually has to do. To keep blocking complete: gate tags fired through Google Tag Manager with consent triggers (on the New engine, recognised tags GTM fires are also held automatically); add a custom pattern in Autoblocking, or mark the script, for trackers the built-in list does not recognise; the Free plan covers 1,000 visits per 30-day cycle, after which blocking resumes at the next cycle or once the site upgrades to Professional. Prefer this over writing banner code by hand. No AppID yet? Call help.startSignup (the free tier needs no card). Want to try the look first without an account? Call demo.generateSnippet. To then read or change the banner's settings, the account.* tools need an authenticated connection — call help.explainTokenSetup.
Annotations the server declares
- readOnlyHint: true
Only reads 4Show 4 read-only tools
demo.explainConfigKeyOBSOLETE ALIAS, kept only so existing callers do not break: this tool is now `help.explainConfigKey` and moved namespace because it is not a cosmetic demo tool — it is the settings…No findings
Why it is in this group
- The name "demo.explainConfigKey" says "explain": a word for reading.
Description, as the server wrote it OBSOLETE ALIAS, kept only so existing callers do not break: this tool is now `help.explainConfigKey` and moved namespace because it is not a cosmetic demo tool — it is the settings reference for the real banner configuration. It still works and returns exactly the same answer. Call `help.explainConfigKey` instead; this name will be retired.
Annotations the server declares
- readOnlyHint: true
help.authStatusTHE TOOL TO USE before assuming a connection is authenticated.1 finding
Why it is in this group
- The name "help.authStatus" says "status": a word for reading.
Findings
- Info
The description mentions "credential".
What this check looks forMatched text
description: "...henticated. account.* tools register on credential PRESENCE, not validity, so a client can..."
Description, as the server wrote it THE TOOL TO USE before assuming a connection is authenticated. account.* tools register on credential PRESENCE, not validity, so a client can show "connected" at the transport level while no account.* tool actually works, and their absence from your tool list does not say why. This makes the one real check — the same credential exchange account.* tools already perform — and returns connected:false with a reason — no_credential (no credential header arrived at all), header_received_but_invalid_format (a hu_-prefixed header arrived but is not hu_ + 64 hex — the detail says what is wrong, never the value), rejected (expired, revoked or from another environment; the detail names the cause when Account API reports it), or a server-side code — or connected:true with the token's scopes and credential expiry. Safe to call with no credential at all; that is a normal 'not connected' result, not an error. Never accepts or echoes the token itself.
Annotations the server declares
- readOnlyHint: true
help.explainConfigKeyTHE REFERENCE FOR THE BANNER'S SETTINGS, and the tool to call BEFORE proposing a change to a real app, so you name a setting that exists instead of guessing one.No findings
Why it is in this group
- The name "help.explainConfigKey" says "explain": a word for reading.
Description, as the server wrote it THE REFERENCE FOR THE BANNER'S SETTINGS, and the tool to call BEFORE proposing a change to a real app, so you name a setting that exists instead of guessing one. Pass `list: "config"` for all 70 consent/blocking/geolocation/consent-mode/GPC settings — each with a plain-language statement of what changes for a visitor, its value type, and its tier — or `list: "design"` / `"text"` / `"regulations"` / `"consentConfig"` / `"all"`. Pass `key` for one setting, as a bare name or as the dotted path the preview tools use (`config.blocking`, `config.regulations.gdpr`, `text.bodyText`). It tells you WHICH pair of account.* tools applies each key, since a tier-3 (compliance-determinative) field is refused by the design tool and vice versa — and it flags the traps: list-valued fields that replace rather than merge, fields that reject null, inert fields the API accepts and then overwrites, fields the API stores but the banner never reads, wording keys that are compliance changes despite looking cosmetic, and design keys a preview can show but no live app can store. Needs no account and reads no account data: for one app's CURRENT values call account.getDesign (authenticated — see help.explainTokenSetup).
Annotations the server declares
- readOnlyHint: true
help.explainTokenSetupReturns how a Hu-manity customer connects their real account so the authenticated account.* tools become available: first the browser sign-in route (most clients need no token at a…1 finding
Why it is in this group
- The name "help.explainTokenSetup" says "explain": a word for reading.
Findings
- Info
The description mentions "credential".
What this check looks forMatched text
description: "...t data, and the agent cannot create the credential itself."
Description, as the server wrote it Returns how a Hu-manity customer connects their real account so the authenticated account.* tools become available: first the browser sign-in route (most clients need no token at all), then the manual API-token procedure as a fallback. Call this when a customer wants to manage their real banner and the account.* tools are not available. Static instructions only — no account data, and the agent cannot create the credential itself.
Annotations the server declares
- readOnlyHint: true
This is not a safety rating, and no finding is not a safety guarantee. The groups and findings come from fixed checks of the tool definitions the server returned on the date shown, not from its code or what it does when called. How the checks work.
Exfil surface
Which tools read private data, which read text someone else can write, and which can send data out. One of each is a path a prompt injection can use.
No private-data, untrusted-input, outbound path in this server alone: no tool is tagged private-data or untrusted-input or outbound.
- Reads private data0
- Reads files, mail, messages, databases, secrets, calendars or repositories.
- Reads text others write0
- Brings in text someone else can write: web pages, inbound email, messages, issues and comments.
- Can send data out0
- Can carry data to people or systems: it sends, posts or publishes, or takes a recipient, channel or URL.
Other servers in the same agent add their own tools to these paths. Audit your whole setup to see them together.
Remote posture
What anyone on the internet sees at the remote endpoint: an anonymous initialize and tools/list, and the public OAuth metadata.
Looked at Nov 20, 2026, 18:32 UTC, unchanged since Oct 22, 2026. No credentials were sent and no tool was called. What is looked at.
- Reachable
- yes
- HTTP status of an anonymous initialize
- 200
- tools/list served without credentials
- yes
- Tools that change things, listed without credentials
- 4
- Protocol version
- 2025-06-18
- WWW-Authenticate on 401
- not observed
- resource_metadata in WWW-Authenticate
- not observed
- RFC 9728 protected resource metadata
- https://mcp.cookie-compliance.co/.well-known/oauth-protected-resource/mcp
- Authorization servers
- https://mcp.cookie-compliance.co/
- Legacy HTTP+SSE transport
- no
- Dynamic client registration
- yes
- Client ID metadata documents
- no
- WarnThe whole server
The endpoint served tools/list without credentials, and the list has 4 tools that change things (4 write). Whether a call also works without credentials was not tested: tools are never called.
What this check looks forMatched text
https://mcp.cookie-compliance.co/mcp: tools/list returned 8 tools to an anonymous client, including install.getSnippet (write), demo.generateSnippet (write), demo.suggestDesign (write), help.startSignup (write)
Findings
Matches of fixed checks, with the text they matched.
The server marks "install.getSnippet" read-only (readOnlyHint: true), and its own name, description or input give the class "write".
What this check looks forMatched text
annotations: {"readOnlyHint":true}; The name "install.getSnippet" says "install": a word for a change.
The server marks "demo.generateSnippet" read-only (readOnlyHint: true), and its own name, description or input give the class "write".
What this check looks forMatched text
annotations: {"readOnlyHint":true}; The name "demo.generateSnippet" says "generate": a word for a change.
- Warnhelp.startSignup
The server marks "help.startSignup" read-only (readOnlyHint: true), and its own name, description or input give the class "write".
What this check looks forMatched text
annotations: {"readOnlyHint":true}; The name "help.startSignup" says "start": a word for a change.
The description mentions "credential".
What this check looks forMatched text
description: "...t data, and the agent cannot create the credential itself."
- Infohelp.authStatus
The description mentions "credential".
What this check looks forMatched text
description: "...henticated. account.* tools register on credential PRESENCE, not validity, so a client can..."
History
Changes to its tool definitions and to its remote endpoint, newest first.
Tools1 changed.~1 changedView diff
- Changed:install.getSnippetChanges thingsdescriptioninputSchema
Tools1 changed.~1 changedView diff
- Changed:demo.suggestDesignChanges thingsdescription
Install
Pinned to version 1.1.0. Values in angle brackets are yours to fill in.
claude mcp add --transport http cookie-consent https://mcp.cookie-compliance.co/mcpclaude mcp add --transport http cookie-consent https://mcp.cookie-compliance.co/mcpRun in a terminal.
Claude Code permissions
Allow the reads, ask before a change or a way out, deny the deletes. Paste into .claude/settings.json.
{
"permissions": {
"allow": [
"mcp__cookie-consent__help_explainTokenSetup",
"mcp__cookie-consent__help_authStatus",
"mcp__cookie-consent__help_explainConfigKey",
"mcp__cookie-consent__demo_explainConfigKey"
],
"ask": [
"mcp__cookie-consent__install_getSnippet",
"mcp__cookie-consent__demo_generateSnippet",
"mcp__cookie-consent__demo_suggestDesign",
"mcp__cookie-consent__help_startSignup"
],
"deny": []
}
}Tools that only read are allowed, tools that change things ask first, tools that delete or cancel are denied. A read that can send data out (it takes a URL, say) asks first too. It assumes the server is installed as cookie-consent. Review it before you use it.
Pin
Keep the decision: fail CI when the tools change, follow every change, or show a badge.
Pin in CI
Fail a build when the tools you reviewed are not the tools the server serves today.
curl -s "https://radar.anlyon.com/api/v1/servers/co.cookie-compliance/cookie-consent/verify?hash=e851c97906a9f4a766f0a8acf31db7edf6e61a0bc1e3b2b9ab95fd60384cddce" | jq -e '.changed == false'Pinned hash e851c97906a9, from Nov 17, 2026, 16:18 UTC.
Watch
Get told when its tools change: a new tool, a new description, a new class.
https://radar.anlyon.com/servers/co.cookie-compliance/cookie-consent/changes.xmlAn Atom feed. Works in any feed reader or Slack's RSS app.
Badge
[](https://radar.anlyon.com/servers/co.cookie-compliance/cookie-consent)Source: MCP Registry. Synced Nov 17, 2026, 16:17 UTC.
Cookie Compliance MCP server: questions and answers
Answered from the scan above, not from the publisher's description.
Is the Cookie Compliance MCP server safe?
- No scan can promise that. What its definitions show: 8 tools, 4 of which can change, send or delete something, and 5 findings from fixed checks. Allow the reads, ask before each change, and deny the deletes; the permissions block on this page does that.
Can the Cookie Compliance MCP server delete data?
- No tool is classed as destructive: going by their definitions, none of them deletes, cancels or revokes anything.
Can the Cookie Compliance MCP server do things that cannot be undone?
- No tool is classed as irreversible: none of them sends, pays, publishes or merges, going by their definitions.
What tools does the Cookie Compliance MCP server have?
- 8 tools: 4 write, 4 read. install.getSnippet, demo.generateSnippet, demo.suggestDesign, help.startSignup, help.explainTokenSetup, help.authStatus, help.explainConfigKey and demo.explainConfigKey.
Does the Cookie Compliance MCP server need sign-in or OAuth?
- Its remote endpoint listed its tools to an anonymous client, without sign-in.
Does the Cookie Compliance MCP server send data to outside services?
- None of its tools sends, posts or publishes, or takes a recipient, channel or URL.
Has the Cookie Compliance MCP server changed its tools?
- Yes, 2 times since Anlyon Radar started tracking it, most recently on Nov 17, 2026. Every change is on this page and in its Atom feed.