Skip to content

Contrie MCP server

com.contrie/contriecontrie.comstreamable-httpv1.2.0Scanned

Extract structured data from a web page, read it as markdown, verify values against the source.

9 tools: 1 can delete or cancel things, 3 change things and 5 only read.

Tools: 1 delete, 0 irreversible, 3 write, 5 read.
  • Reads private data · 1
  • Reads text others write · 6
  • Can send data out · 4
  • 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 Contrie MCP server can do

The Contrie MCP server is published by contrie.com, a domain the MCP Registry verified. It has 9 tools. It can delete or cancel things (1 tool: contrie_unwatch) and create or change data (3 tools: contrie_extract, contrie_verify and contrie_watch). 5 tools only read.

1 tool reads private data (contrie_read); 6 tools bring in text other people can write (contrie_search, contrie_extract, contrie_read and 3 more); 4 tools can send data out of your system (contrie_extract, contrie_read, contrie_verify and 1 more). On its own it has all three parts of a prompt-injection exfiltration path: private data, text others write, and a way out.

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, and its tools changed once since Anlyon Radar first saw them.

Last scanned Nov 17, 2026, version 1.2.0.

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.

9 tools

Deletes or cancels 1Claude Code: deny
  • contrie_unwatchStop one of your monitors so it makes no further scheduled runs and no further charges; a run already in progress can still finish and be charged.No findings

    Why it is in this group

    • The description's first verb is "stop": a word for a change.
    • The server marks it destructive (destructiveHint: true).
    Description, as the server wrote it
    Stop one of your monitors so it makes no further scheduled runs and no further charges; a run already in progress can still finish and be charged. History stays readable. Returns the monitor with active: false; stopping an already stopped monitor returns it unchanged. A stopped monitor cannot be restarted. Charges no credits. Requires authentication: sign in with your Contrie account (OAuth), or send an API key in the Authorization header.

    Annotations the server declares

    • title: Stop a monitor
    • readOnlyHint: false
    • destructiveHint: true
    • idempotentHint: true
    • openWorldHint: false
Changes things 3Claude Code: ask
  • contrie_extractExtract structured data from one public web page.Reads text others writeCan send data outNo 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.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    • outboundThe input takes a URL (input.url): a request to a URL can carry data to whoever runs it.
    Description, as the server wrote it
    Extract structured data from one public web page. Describe the fields you want in natural language and/or pass a JSON Schema for the exact shape. Returns JSON plus metadata: qualityScore (0-100; structured acceptance also requires success and valid), grounding (value presence, not field attribution), high-level progress and credits charged. Also returns groundingFields, a per-field grounding verdict with page excerpts, so you can see which extracted values have matching source text; this does not establish factual truth or correct field association. metadata.binding, when present, lists the paths of fields whose value is on the page while the passages about that field state a different value. Requires authentication: sign in with your Contrie account (OAuth), or send an API key in the Authorization header. A small set of recorded sample requests works without either. An accepted answer costs 1, 3, 8 or 15 credits; a refused one costs 0. Each plan caps one request's credits, rendering included (Free 3, Builder 5, Pro 17); an answer above the cap is refused and costs 0. Not a crawler or a search engine: one page per call. Eligible JavaScript shells can use bounded browser rendering; successful rendered answers add 2 credits.

    Annotations the server declares

    • title: Extract structured data from a web page
    • readOnlyHint: true
    • openWorldHint: true
  • contrie_verifyCheck data you already have against the page it claims to come from, field by field.Reads text others writeCan send data outNo 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.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    • outboundThe input takes a URL (input.url): a request to a URL can carry data to whoever runs it.
    Description, as the server wrote it
    Check data you already have against the page it claims to come from, field by field. Takes data from any source plus a url (or raw html/text) and reports which values actually appear in the source and which do not, with a supporting excerpt for each. This check charges no credits. Requires authentication: sign in with your Contrie account (OAuth), or send an API key in the Authorization header. It answers "is this value present in this page", NOT "does the page say this about that": a value that is correct for a different field still reads as grounded.

    Annotations the server declares

    • title: Verify extracted data against its source page
    • readOnlyHint: true
    • openWorldHint: true
  • contrie_watchWatch one public web page for one question and be told when the answer changes.Reads text others writeCan send data outNo findings

    Why it is in this group

    • Nothing in its name or description says what it does, so it is counted as a write.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    • outboundThe input takes a URL (input.url): a request to a URL can carry data to whoever runs it.
    Description, as the server wrote it
    Watch one public web page for one question and be told when the answer changes. Runs the extraction now as a baseline, then on a schedule with `every` minutes between slots (minimum 15); delayed or missed slots can occur. On a change Contrie records the before and after responses and their evidence status plus a field diff in the monitor's history, and attempts one POST to webhookUrl if given (signed X-Contrie-Signature; the secret is returned once). Every run with an accepted answer, the baseline included, is charged like one extraction (1, 3, 8 or 15 credits, plus 2 when rendered), so an hourly monitor can make up to 24 charged runs a day until it is stopped. Plans cap active monitors (Free 3, Builder 25, Pro 100). Creation is not idempotent: if a response is lost, list your monitors before retrying, or the retry creates a second monitor that also charges. Requires authentication: sign in with your Contrie account (OAuth), or send an API key in the Authorization header.

    Annotations the server declares

    • title: Watch a page for changes to an answer
    • readOnlyHint: false
    • destructiveHint: false
    • idempotentHint: false
    • openWorldHint: true
Only reads 5Show 5 read-only tools
  • contrie_accountRead your own account’s used, reserved and unreserved included credits, next allowance reset and current restrictions.No findings

    Why it is in this group

    • The description's first verb is "read": a word for reading.
    Description, as the server wrote it
    Read your own account’s used, reserved and unreserved included credits, next allowance reset and current restrictions. Costs zero extraction credits and reserves no capacity. Eligibility is advisory: source, service and request-specific checks still apply. Requires account authorization.

    Annotations the server declares

    • title: Check account credits and current restrictions
    • readOnlyHint: true
    • openWorldHint: false
  • contrie_changesRead a monitor's history: its latest answer with evidence status and the last 20 events (baseline, change, error).No findings

    Why it is in this group

    • The description's first verb is "read": a word for reading.
    Description, as the server wrote it
    Read a monitor's history: its latest answer with evidence status and the last 20 events (baseline, change, error). Pass nextCursor as before for older events, or eventId to recover one exact event; do not combine them. Also returns monitor.runStatus for one current original slot: scheduled (future with no outstanding work), pending (work may still complete or publish), needs_review (reason result_unconfirmed or result_unavailable), or unknown (ambiguous evidence, including a recent stopped claim whose request ownership cannot be proved). Each object includes scheduledFor, the UTC millisecond due time; only needs_review includes reason. Stopped idle runStatus is null only with complete positive no-outstanding current-slot evidence. statusRead {coverage: complete, observedAt} covers the monitor/status database statement observation, not separately read events or fleet/history counts. Pending promises no running process, charge, completion time or recovery; review reasons imply neither refund, failed extraction nor semantic correctness. Snapshot failure is SERVICE_UNAVAILABLE. Reading does not change monitoring or accounting state; ordinary authentication and rate-limit effects apply. Every change carries the full before and after responses with evidence, and a diff of the fields that moved. Charges no credits. Requires authentication: sign in with your Contrie account (OAuth), or send an API key in the Authorization header.

    Annotations the server declares

    • title: Read what a monitor has seen
    • readOnlyHint: true
    • openWorldHint: false
  • contrie_monitorsList your active monitors: id, url, question, cadence, next scheduled run, latest answer and current-slot runStatus.Reads text others writeNo findings

    Why it is in this group

    • The description's first verb is "list": a word for reading.

    What it can reach

    • untrusted-inputThe description's first sentence says "list" and "url": it reads web pages, which anyone can write.
    Description, as the server wrote it
    List your active monitors: id, url, question, cadence, next scheduled run, latest answer and current-slot runStatus. Includes one statusRead for the complete account snapshot, even when empty; unavailable snapshots fail. Stopped monitors are not listed; their history stays readable. Status grants no recovery work or financial conclusion. Charges no credits. Requires authentication: sign in with your Contrie account (OAuth), or send an API key in the Authorization header.

    Annotations the server declares

    • title: List your active monitors
    • readOnlyHint: true
    • openWorldHint: false
  • contrie_readRead one public web page as clean, structure-preserving markdown (headings, lists, tables, links) with common page boilerplate removed where recognized.Reads private dataReads text others writeCan send data outNo findings

    Why it is in this group

    • The name "contrie_read" says "read": a word for reading.

    What it can reach

    • private-dataThe description's first sentence says "read" and "table": it reads databases.
    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    • untrusted-inputThe description's first sentence says "read" and "web": it reads web pages, which anyone can write.
    • untrusted-inputThe input takes a URL (input.url), and the tool reads it: it reads web pages, which anyone can write.
    • outboundThe input takes a URL (input.url): a request to a URL can carry data to whoever runs it.

    Claude Code: ask, because it is tagged outbound.

    Description, as the server wrote it
    Read one public web page as clean, structure-preserving markdown (headings, lists, tables, links) with common page boilerplate removed where recognized. A successful read costs 1 credit; a failed one costs 0. Returns the full REST-shaped JSON response in this tool's text content: data is null, markdown holds the page, and metadata includes credits and evidence status. Returns the page as markdown, not typed fields. Contrie reads at most the first 100,000 characters of the page's markdown; metadata.evidenceCoverage.sourceTruncated is true when it cut. For a long page, send maxChars to get verbatim passages inside that size (whole blocks where they fit; a passage larger than the size is cut and metadata.focus.cutInsideBlock is true), and objective to get the passages that share the most words with what you are looking for; metadata.focus lists what was left out and where to continue. Requires authentication: sign in with your Contrie account (OAuth), or send an API key in the Authorization header. Eligible JavaScript shells can use bounded browser rendering; successful rendered reads add 2 credits.

    Annotations the server declares

    • title: Read a web page as markdown
    • readOnlyHint: true
    • openWorldHint: 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.

24 paths in this server alone from private data, through text someone else wrote, to a way out. Denying 1 private-data tool breaks every one.

Reads private data1
Reads files, mail, messages, databases, secrets, calendars or repositories.
Reads text others write6
Brings in text someone else can write: web pages, inbound email, messages, issues and comments.
Can send data out4
Can carry data to people or systems: it sends, posts or publishes, or takes a recipient, channel or URL.
deny to break every path
{
  "permissions": {
    "deny": [
      "mcp__contrie__contrie_read"
    ]
  }
}

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, 18:36 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://www.contrie.com/.well-known/oauth-protected-resource/mcp
Authorization servers
https://clerk.contrie.com
Legacy HTTP+SSE transport
no
Dynamic client registration
no
Client ID metadata documents
yes
  • HighThe whole server

    The endpoint served tools/list without credentials, and the list has 4 tools that change things (1 destructive, 3 write). Whether a call also works without credentials was not tested: tools are never called.

    Matched text

    https://www.contrie.com/mcp: tools/list returned 9 tools to an anonymous client, including contrie_unwatch (destructive), contrie_extract (write), contrie_verify (write), contrie_watch (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.

  1. Tools6 changed.~6 changedView diff

Install

Pinned to version 1.2.0. Values in angle brackets are yours to fill in.

claude mcp add --transport http contrie https://www.contrie.com/mcp
claude mcp add --transport http contrie https://www.contrie.com/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__contrie__contrie_search",
      "mcp__contrie__contrie_account",
      "mcp__contrie__contrie_changes",
      "mcp__contrie__contrie_monitors"
    ],
    "ask": [
      "mcp__contrie__contrie_extract",
      "mcp__contrie__contrie_read",
      "mcp__contrie__contrie_verify",
      "mcp__contrie__contrie_watch"
    ],
    "deny": [
      "mcp__contrie__contrie_unwatch"
    ]
  }
}

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 contrie. Review it before you use it.

Capability tags behind 6 rules
  • mcp__contrie__contrie_search untrusted-input
  • mcp__contrie__contrie_extract untrusted-input, outbound
  • mcp__contrie__contrie_read private-data, untrusted-input, outbound
  • mcp__contrie__contrie_verify untrusted-input, outbound
  • mcp__contrie__contrie_watch untrusted-input, outbound
  • mcp__contrie__contrie_monitors 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/com.contrie/contrie/verify?hash=fb58d45e57bafc93976e628b0c167fb2457dbdb84fd4559754ea5116f0566379" | jq -e '.changed == false'

Pinned hash fb58d45e57ba, from Nov 17, 2026, 20:18 UTC.

Watch

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

https://radar.anlyon.com/servers/com.contrie/contrie/changes.xml

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

Badge

Anlyon Radar badge for com.contrie/contrie
[![Anlyon Radar](https://radar.anlyon.com/badge/com.contrie/contrie.svg)](https://radar.anlyon.com/servers/com.contrie/contrie)

Source: MCP Registry. Synced Nov 17, 2026, 20:17 UTC.

Contrie MCP server: questions and answers

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

Is the Contrie MCP server safe?

No scan can promise that. What its definitions show: 9 tools, 4 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 Contrie MCP server delete data?

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

Can the Contrie 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 Contrie MCP server have?

9 tools: 1 destructive, 3 write, 5 read. contrie_unwatch, contrie_extract, contrie_verify, contrie_watch, contrie_search, contrie_read, contrie_account, contrie_changes and 1 more.

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

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

Does the Contrie MCP server send data to outside services?

4 tools can carry data out (it sends, posts or publishes, or takes a recipient, channel or URL): contrie_extract, contrie_read, contrie_verify and contrie_watch.

Has the Contrie MCP server changed its tools?

Yes, once since Anlyon Radar started tracking it, most recently on Nov 17, 2026. Every change is on this page and in its Atom feed.