AgentPixel MCP server
io.agentpixel/agentpixelagentpixel.iostreamable-httpv0.1.4Scanned
Most visitors never fill out a form. Turn the ones who click Accept into contacts your agent reads.
14 tools: 1 can delete or cancel things, 1 do something that can't be undone, 6 change things and 6 only read.
- Reads private data · 1
- Reads text others write · 6
- Can send data out · 1
- No findings
Tools listed without sign-in. Remote posture
With Anlyon
1 tool here can delete or cancel things and 1 does something that can't be undone. 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 AgentPixel MCP server can do
The AgentPixel MCP server is published by agentpixel.io, a domain the MCP Registry verified. It has 14 tools. It can delete or cancel things (1 tool: configure_consent), do things that cannot be undone, such as sending, paying or publishing (resend_code) and create or change data (6 tools: try_sample, signup_free, confirm_signup, create_api_key and 2 more). 6 tools only read.
1 tool reads private data (get_contacts); 6 tools bring in text other people can write (signup_free, resend_code, create_pixel and 3 more); 1 tool can send data out of your system (resend_code). 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.
Last scanned Nov 11, 2026, version 0.1.4.
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.
14 tools
Deletes or cancels 1Claude Code: deny
configure_consentTurn on and set up the consent bar for a pixel.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 destructive (destructiveHint: true).
Description, as the server wrote it Turn on and set up the consent bar for a pixel. Changing any setting makes a new policy version, and returning visitors are asked again. Returns the policy link and the script tag to put on the site.
Annotations the server declares
- title: Configure consent and policy
- readOnlyHint: false
- destructiveHint: true
- idempotentHint: false
- openWorldHint: false
Can't be undone 1Claude Code: ask
resend_codeSend a new 6-digit code to the email of an existing account.Reads text others writeCan send data outNo findings
Why it is in this group
- The name "resend_code" says "resend": a word for sending, and what is sent cannot be taken back.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
- outboundThe name "resend_code" says "resend": it sends or publishes something to other people.
Description, as the server wrote it Send a new 6-digit code to the email of an existing account. Use it when the code expired, was lost or was entered wrong 5 times. Limits: 1 per minute, 5 per day. The old code stops working. With an expired signup_token, use the new code on POST /v1/accounts/{account_id}/signup-token to get a new token.
Annotations the server declares
- title: Send a new email code
- readOnlyHint: false
- destructiveHint: false
- idempotentHint: false
- openWorldHint: true
Changes things 6Claude Code: ask
check_installFetch the pixel's homepage and store whether the tag is there.Reads text others writeNo findings
Why it is in this group
- The name "check_install" says "install": a word for a change.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
Description, as the server wrote it Fetch the pixel's homepage and store whether the tag is there. install_status becomes installed or not_installed. check_error says why a check could not decide. Evaluator accounts fetch only the fixed https://staging.allsourcedata.io/evaluation-test/ page and reject redirects.
Annotations the server declares
- title: Check and save installation
- readOnlyHint: false
- destructiveHint: false
- idempotentHint: false
- openWorldHint: true
confirm_signupConfirm the email with the 6-digit code, or check whether the email link was clicked.No findings
Why it is in this group
- The input takes a statement or command to run (input.code), which can change anything it can reach.
Description, as the server wrote it Confirm the email with the 6-digit code, or check whether the email link was clicked. Preferred: confirm_signup(account_id, code, signup_token). The first confirmation returns api_key; store it, it is shown once. Link clicked instead: call confirm_signup(account_id, signup_token) every 5 seconds, up to 10 minutes, until email_verified is true, then call create_api_key. A code allows 5 attempts and lasts 24 hours. For a new code, call resend_code(email).
Annotations the server declares
- title: Confirm signup email
- readOnlyHint: false
- destructiveHint: false
- idempotentHint: true
- openWorldHint: false
create_api_keyIssue an API key after the email is confirmed.No findings
Why it is in this group
- The name "create_api_key" says "create": a word for a change.
Description, as the server wrote it Issue an API key after the email is confirmed. Not needed when confirm_signup already returned api_key. Each call issues another key, so call it once. Returns the key once. Store it securely and attach it to the connection.
Annotations the server declares
- title: Issue private API key
- readOnlyHint: false
- destructiveHint: false
- idempotentHint: false
- openWorldHint: false
create_pixelCreate a pixel for a domain.Reads text others writeNo findings
Why it is in this group
- The name "create_pixel" says "create": a word for a change.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
Description, as the server wrote it Create a pixel for a domain. Does not create an API key. The account must already be active. Next: call configure_consent(pixel_id, controller_name, purposes) and install the snippet it returns. Do not install a snippet before consent is configured. install_status starts as not_checked until check_install runs.
Annotations the server declares
- title: Create site pixel
- readOnlyHint: false
- destructiveHint: false
- idempotentHint: false
- openWorldHint: true
signup_freeStart a 30-day trial, no card.Reads text others writeNo findings
Why it is in this group
- The description's first verb is "start": a word for a change.
- The input takes a statement or command to run (input.code), which can change anything it can reach.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
Description, as the server wrote it Start a 30-day trial, no card. After that, $1 per account per month for 3 pixels and 500 new contacts per UTC calendar month. Sends a 6-digit code to the email address. Returns account_id and signup_token; keep both. Next: ask the person for the 6-digit code in the AgentPixel email, then call confirm_signup(account_id, code, signup_token). The signup_token lasts 24 hours.
Annotations the server declares
- title: Start 30-day trial
- readOnlyHint: false
- destructiveHint: false
- idempotentHint: false
- openWorldHint: true
try_sampleLabeled sample of what an account gets: a pixel, a consent receipt and two contacts.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 Labeled sample of what an account gets: a pixel, a consent receipt and two contacts. Needs no account and no key. Every record is synthetic and marked so. Call this first to show the person what AgentPixel returns, then call signup_free.
Annotations the server declares
- title: See sample output
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: false
Only reads 6Show 6 read-only tools
get_contactsContacts for one pixel.Reads private dataNo findings
Why it is in this group
- The name "get_contacts" says "get": a word for reading.
What it can reach
- private-dataThe name "get_contacts" says "get" and "contact": it reads contacts.
- private-dataThe description's first sentence says "contact": it reads contacts.
Description, as the server wrote it Contacts for one pixel. Pass on as a UTC day, or after as the high_water from the previous pull. Use high_water as the next cursor. New contacts can arrive between reads. This does not identify new visitors. Evaluation returns labeled synthetic examples only after current resolution consent is saved; withdrawal hides them.
Annotations the server declares
- title: Read contacts
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: false
get_pixelOne pixel, including the HTML snippet to paste on its site.Reads text others writeNo findings
Why it is in this group
- The name "get_pixel" says "get": a word for reading.
What it can reach
- untrusted-inputThe description's first sentence says "html": it reads web pages, which anyone can write.
Description, as the server wrote it One pixel, including the HTML snippet to paste on its site.
Annotations the server declares
- title: Read pixel and install tag
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: false
get_policyHosted privacy and data-use policy HTML for a pixel.Reads text others writeNo findings
Why it is in this group
- The name "get_policy" says "get": a word for reading.
What it can reach
- untrusted-inputThe description's first sentence says "html": it reads web pages, which anyone can write.
Description, as the server wrote it Hosted privacy and data-use policy HTML for a pixel. A pixel with consent turned off has no policy yet.
Annotations the server declares
- title: Read hosted privacy notice
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: false
list_pixelsPixels on this account.No findings
Why it is in this group
- The name "list_pixels" says "list": a word for reading.
Description, as the server wrote it Pixels on this account. Each item has pixel_id, domain, and install_status.
Annotations the server declares
- title: List my pixels
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: false
setup_statusWhether this account can use pixels and whether a tag is installed.No findings
Why it is in this group
- The name "setup_status" says "status": a word for reading.
Description, as the server wrote it Whether this account can use pixels and whether a tag is installed. setup_complete is true once at least one pixel's tag has been seen on its site. good_standing is the account, not the tag.
Annotations the server declares
- title: Read installation progress
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: false
whoamiWhich account this connection's API key belongs to.No findings
Why it is in this group
- The name "whoami" says "whoami": a word for reading.
Description, as the server wrote it Which account this connection's API key belongs to. status is the account state after email confirmation. plan is "free" (the trial plan id) during the 30-day trial, paid after the subscription, or unlimited for internal accounts. trial_ends_at is when the trial ends. key_prefix is the start of the key, not the key.
Annotations the server declares
- title: Check account identity
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: false
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.
6 paths in this server alone from private data, through text someone else wrote, to a way out. Denying 1 outbound 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 out1
- Can carry data to people or systems: it sends, posts or publishes, or takes a recipient, channel or URL.
{
"permissions": {
"deny": [
"mcp__agentpixel__resend_code"
]
}
}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, 18:55 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
- 8
- Protocol version
- 2025-06-18
- WWW-Authenticate on 401
- not observed
- resource_metadata in WWW-Authenticate
- not observed
- RFC 9728 protected resource metadata
- https://agentpixel.io/.well-known/oauth-protected-resource/mcp
- Authorization servers
- https://agentpixel.io
- Legacy HTTP+SSE transport
- no
- Dynamic client registration
- yes
- Client ID metadata documents
- no
- HighThe whole server
The endpoint served tools/list without credentials, and the list has 8 tools that change things (1 destructive, 1 irreversible, 6 write). Whether a call also works without credentials was not tested: tools are never called.
What this check looks forMatched text
https://agentpixel.io/mcp: tools/list returned 14 tools to an anonymous client, including configure_consent (destructive), resend_code (irreversible), try_sample (write), signup_free (write), confirm_signup (write) and 3 more
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 0.1.4. Values in angle brackets are yours to fill in.
claude mcp add --transport http agentpixel https://agentpixel.io/mcpclaude mcp add --transport http agentpixel https://agentpixel.io/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__agentpixel__whoami",
"mcp__agentpixel__setup_status",
"mcp__agentpixel__list_pixels",
"mcp__agentpixel__get_pixel",
"mcp__agentpixel__get_policy",
"mcp__agentpixel__get_contacts"
],
"ask": [
"mcp__agentpixel__try_sample",
"mcp__agentpixel__signup_free",
"mcp__agentpixel__confirm_signup",
"mcp__agentpixel__resend_code",
"mcp__agentpixel__create_api_key",
"mcp__agentpixel__create_pixel",
"mcp__agentpixel__check_install"
],
"deny": [
"mcp__agentpixel__configure_consent"
]
}
}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 agentpixel. Review it before you use it.
Capability tags behind 7 rules
- mcp__agentpixel__signup_free untrusted-input
- mcp__agentpixel__resend_code untrusted-input, outbound
- mcp__agentpixel__create_pixel untrusted-input
- mcp__agentpixel__get_pixel untrusted-input
- mcp__agentpixel__check_install untrusted-input
- mcp__agentpixel__get_policy untrusted-input
- mcp__agentpixel__get_contacts private-data
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.agentpixel/agentpixel/verify?hash=80fc3c293f6a72ea0e250cf0d4f871369118ff464c50a6791105b7c0e831229e" | jq -e '.changed == false'Pinned hash 80fc3c293f6a, from Nov 11, 2026, 15:17 UTC.
Watch
Get told when its tools change: a new tool, a new description, a new class.
https://radar.anlyon.com/servers/io.agentpixel/agentpixel/changes.xmlAn Atom feed. Works in any feed reader or Slack's RSS app.
Badge
[](https://radar.anlyon.com/servers/io.agentpixel/agentpixel)Source: MCP Registry · repository. Synced Oct 21, 2026, 18:13 UTC.
AgentPixel MCP server: questions and answers
Answered from the scan above, not from the publisher's description.
Is the AgentPixel MCP server safe?
- No scan can promise that. What its definitions show: 14 tools, 8 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 AgentPixel MCP server delete data?
- Yes. 1 tool can delete, cancel or revoke something: configure_consent.
Can the AgentPixel MCP server do things that cannot be undone?
- Yes. 1 tool can send, pay, publish or merge: resend_code.
What tools does the AgentPixel MCP server have?
- 14 tools: 1 destructive, 1 irreversible, 6 write, 6 read. configure_consent, resend_code, try_sample, signup_free, confirm_signup, create_api_key, create_pixel, check_install and 6 more.
Does the AgentPixel MCP server need sign-in or OAuth?
- Its remote endpoint listed its tools to an anonymous client, without sign-in.
Does the AgentPixel MCP server send data to outside services?
- 1 tool can carry data out (it sends, posts or publishes, or takes a recipient, channel or URL): resend_code.
Has the AgentPixel MCP server changed its tools?
- Not since Anlyon Radar started tracking it. Last scanned Nov 11, 2026.