Skip to content

LeakData MCP server

io.github.leakdatahq/leakdata-mcpleakdatahqstreamable-httpv1.3.2Scanned

Check breach exposure for your verified email and check locally computed password hash prefixes.

7 tools: 1 can delete or cancel things, 4 change things and 2 only read.

Tools: 1 delete, 0 irreversible, 4 write, 2 read.
  • Reads private data · 1
  • Reads text others write · 2
  • Can send data out · 0
  • No findings

Tools listed without sign-in. Remote posture

With Anlyon

1 tool here can delete or cancel things. Govern it.

  • Anlyon lists the tools and adopts none.
  • Each tool you adopt gets approvals and receipts.
  • Your agent never holds the key.
Govern this server →

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 LeakData MCP server can do

The LeakData MCP server is published by the GitHub account leakdatahq. It has 7 tools. It can delete or cancel things (1 tool: leakdata.mark_step_done) and create or change data (4 tools: leakdata.protection_summary, leakdata.recent_breaches, leakdata.company_monitoring and leakdata.action_plan). 2 tools only read.

1 tool reads private data (leakdata.search); 2 tools bring in text other people can write (leakdata.search and leakdata.breach_lookup).

None of the fixed checks matched anything in its tool definitions. That is not a safety guarantee.

Its remote endpoint lists its tools without sign-in.

Last scanned Nov 21, 2026, version 1.3.2.

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.

7 tools

Deletes or cancels 1Claude Code: deny
  • leakdata.mark_step_doneUse this only after the user says they completed (or want to reopen) a specific step from leakdata.action_plan, for example "I changed my Canva password".No findings

    Why it is in this group

    • The name "leakdata.mark_step_done" says "mark": a word for a change.
    • The server marks it destructive (destructiveHint: true).
    Description, as the server wrote it
    Use this only after the user says they completed (or want to reopen) a specific step from leakdata.action_plan, for example "I changed my Canva password". Pass that step's stepId from leakdata.action_plan and done=true, or done=false to reopen it; if no action plan result in this conversation contains the step, call leakdata.action_plan first. Saves the change to the user's LeakData action plan; it does not change anything on the breached site. A request to mark all or several steps done is not confirmation: ask which steps the user actually finished and call this once per confirmed step. Never guess a stepId.

    Annotations the server declares

    • title: Mark a step done
    • readOnlyHint: false
    • destructiveHint: true
    • idempotentHint: true
    • openWorldHint: false
Changes things 4Claude Code: ask
  • leakdata.action_planUse this when the linked LeakData user asks what they should do about their breaches, what is left on their list, or wants to be walked through the steps.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
    Use this when the linked LeakData user asks what they should do about their breaches, what is left on their list, or wants to be walked through the steps. Returns the open and done steps of the user's own action plan (for example "Change your password" on a breached site), each with a stepId, the service it applies to, the masked monitored address and whether it is done. Walk the user through open steps one at a time, in the order returned (most urgent first). Report only the returned steps; never invent steps or mark progress yourself. Takes no arguments.

    Annotations the server declares

    • title: Your action plan
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: false
  • leakdata.company_monitoringUse this when someone asks about their own company or employer: whether it had a data breach, whether its domain, staff email addresses or work accounts appear in data breaches, or…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
    Use this when someone asks about their own company or employer: whether it had a data breach, whether its domain, staff email addresses or work accounts appear in data breaches, or how to monitor a business domain. Explains what LeakData checks for a domain the company verifies it owns and returns any public catalog breaches listed for exactly that domain with their details. Explain setup using the returned checks and requirements; do not offer plans, prices or sales links. No sign-in is needed. It does not enable monitoring, search employee addresses, or reveal any person's data; never claim it did. Not listed means not in LeakData's catalog, not that the company was never breached.

    Annotations the server declares

    • title: Understand company monitoring
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: false
  • leakdata.protection_summaryUse this when the linked LeakData user asks how their monitoring is going, what LeakData has found for them, or whether anything new turned up.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
    Use this when the linked LeakData user asks how their monitoring is going, what LeakData has found for them, or whether anything new turned up. Returns the user's own monitored email addresses (masked), when each was last checked, how many findings each has, new findings in the last 30 days, and a count of open action-plan steps. It reports stored monitoring results; for a fresh check of the user's email use leakdata.search, and for the steps themselves use leakdata.action_plan. Takes no arguments. Report only the returned values.

    Annotations the server declares

    • title: Review your monitoring
    • readOnlyHint: true
    • destructiveHint: false
    • openWorldHint: false
    • idempotentHint: true
  • leakdata.recent_breachesUse this when someone asks which data breaches happened recently or what the latest breaches are.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
    Use this when someone asks which data breaches happened recently or what the latest breaches are. Returns the newest breaches in LeakData's public catalog, newest incident date first, with date, affected account count, exposed data types and first steps. No sign-in is needed. It has no date filter: check the returned dates and say so when none fall in the period the user asked about. It is LeakData's catalog, not every breach worldwide. Not for general security, ransomware or vulnerability news.

    Annotations the server declares

    • title: Latest data breaches
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: false
Only reads 2Show 2 read-only tools
  • leakdata.breach_lookupUse this when someone asks whether a company, app or website was hacked or had a data breach, what data leaked, whether a breach notice they received is real, or what to do about a…Reads text others writeNo findings

    Why it is in this group

    • The name "leakdata.breach_lookup" says "lookup": a word for reading.

    What it can reach

    • untrusted-inputThe description's first sentence says "website": it reads web pages, which anyone can write.
    Description, as the server wrote it
    Use this when someone asks whether a company, app or website was hacked or had a data breach, what data leaked, whether a breach notice they received is real, or what to do about a breach they heard of ("Was Canva hacked?", "What leaked in the Ticketmaster breach?"). Searches LeakData's public breach catalog by company, service or domain name and returns each matching breach's date, affected account count, exposed data types, a short summary and the first steps to take. No sign-in is needed. It does not tell whether a particular person's email is in the breach: when emailCheckAvailable is true, offer to check the user's own email with leakdata.search; when it is false, say LeakData cannot check addresses against that breach. When match is "similar", say the name only resembles what the user asked about. No match means the breach is not in LeakData's catalog, not that the company was never breached. If the company is the user's own business or employer, use leakdata.company_monitoring instead. Not for general security news or a person's name.

    Annotations the server declares

    • title: Look up a data breach
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: false
  • leakdata.searchUse this when the linked LeakData user asks whether their own email was in a data breach ("Has my email been leaked?", "Am I in the Canva breach?", "Am I affected by any recent lea…Reads private dataReads text others writeNo findings

    Why it is in this group

    • The name "leakdata.search" says "search": a word for reading.

    What it can reach

    • private-dataThe description's first sentence says "email": it reads mail.
    • untrusted-inputThe description's first sentence says "email": it reads inbound email, which other people can write.
    Description, as the server wrote it
    Use this when the linked LeakData user asks whether their own email was in a data breach ("Has my email been leaked?", "Am I in the Canva breach?", "Am I affected by any recent leaks?"). For a current check, call this tool and report only its returned results; never present earlier results as a new check, and if the call fails say the check could not be completed. For their own email, set type to email and omit query: the server uses the verified primary email of the linked account. If the user typed an address of their own, put it in query; only the verified primary email is checked, and for another address they own the result explains how to add it in LeakData. Never call this for someone else's address or to look up passwords, and never guess an address. No results means no LeakData record matched: never say the email is safe or was never leaked; a breach whose records LeakData does not hold cannot show up. Results for the user's other monitored addresses come from leakdata.protection_summary; company domains use leakdata.company_monitoring.

    Annotations the server declares

    • title: Check your email
    • readOnlyHint: true
    • destructiveHint: false
    • openWorldHint: false
    • idempotentHint: 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 outbound.

Reads private data1
Reads files, mail, messages, databases, secrets, calendars or repositories.
Reads text others write2
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.

Tools listed without sign-in

Looked at Nov 20, 2026, 19:07 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
5
Protocol version
2025-06-18
WWW-Authenticate on 401
not observed
resource_metadata in WWW-Authenticate
not observed
RFC 9728 protected resource metadata
https://leakdata.io/.well-known/oauth-protected-resource/mcp
Authorization servers
https://leakdata.io
Legacy HTTP+SSE transport
no
Dynamic client registration
yes
Client ID metadata documents
yes
  • HighThe whole server

    The endpoint served tools/list without credentials, and the list has 5 tools that change things (1 destructive, 4 write). Whether a call also works without credentials was not tested: tools are never called.

    Matched text

    https://leakdata.io/mcp: tools/list returned 7 tools to an anonymous client, including leakdata.mark_step_done (destructive), leakdata.protection_summary (write), leakdata.recent_breaches (write), leakdata.company_monitoring (write), leakdata.action_plan (write)

    What this check looks for

Findings

Matches of fixed checks, with the text they matched.

No check matched anything in these tool definitions. That is not a safety guarantee.

History

Changes to its tool definitions and to its remote endpoint, newest first.

No change seen yet. Its tools are rescanned every day, and a remote endpoint is looked at once a week.

Install

Pinned to version 1.3.2. Values in angle brackets are yours to fill in.

claude mcp add --transport http leakdata https://leakdata.io/mcp
claude mcp add --transport http leakdata https://leakdata.io/mcp

Run 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.

.claude/settings.json
{
  "permissions": {
    "allow": [
      "mcp__leakdata__leakdata_search",
      "mcp__leakdata__leakdata_breach_lookup"
    ],
    "ask": [
      "mcp__leakdata__leakdata_protection_summary",
      "mcp__leakdata__leakdata_recent_breaches",
      "mcp__leakdata__leakdata_company_monitoring",
      "mcp__leakdata__leakdata_action_plan"
    ],
    "deny": [
      "mcp__leakdata__leakdata_mark_step_done"
    ]
  }
}

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 leakdata. Review it before you use it.

Capability tags behind 2 rules
  • mcp__leakdata__leakdata_search private-data, untrusted-input
  • mcp__leakdata__leakdata_breach_lookup untrusted-input

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/io.github.leakdatahq/leakdata-mcp/verify?hash=0ba3233671921b85ec08c4303b6ec1db688de18c0ac17f4067cade79386223d8" | jq -e '.changed == false'

Pinned hash 0ba323367192, from Nov 21, 2026, 00:17 UTC.

Watch

Get told when its tools change: a new tool, a new description, a new class.

https://radar.anlyon.com/servers/io.github.leakdatahq/leakdata-mcp/changes.xml

An Atom feed. Works in any feed reader or Slack's RSS app.

Badge

Anlyon Radar badge for io.github.leakdatahq/leakdata-mcp
[![Anlyon Radar](https://radar.anlyon.com/badge/io.github.leakdatahq/leakdata-mcp.svg)](https://radar.anlyon.com/servers/io.github.leakdatahq/leakdata-mcp)

Source: MCP Registry. Synced Oct 21, 2026, 18:38 UTC.

LeakData MCP server: questions and answers

Answered from the scan above, not from the publisher's description.

Is the LeakData MCP server safe?

No scan can promise that. What its definitions show: 7 tools, 5 of which can change, send or delete something, and 0 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 LeakData MCP server delete data?

Yes. 1 tool can delete, cancel or revoke something: leakdata.mark_step_done.

Can the LeakData 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 LeakData MCP server have?

7 tools: 1 destructive, 4 write, 2 read. leakdata.mark_step_done, leakdata.protection_summary, leakdata.recent_breaches, leakdata.company_monitoring, leakdata.action_plan, leakdata.search and leakdata.breach_lookup.

Does the LeakData MCP server need sign-in or OAuth?

Its remote endpoint listed its tools to an anonymous client, without sign-in.

Does the LeakData MCP server send data to outside services?

None of its tools sends, posts or publishes, or takes a recipient, channel or URL.

Has the LeakData MCP server changed its tools?

Not since Anlyon Radar started tracking it. Last scanned Nov 21, 2026.