AI Agent Safety Foundational Certificate MCP server
dev.sarnai/certifysarnai.devstreamable-httpv1.0.1Scanned
AI Agent Safety Foundational Certificate: a wallet-signed 40-task exam. First 3 attempts free.
15 tools: 1 can delete or cancel things, 9 change things and 5 only read.
- Reads private data · 0
- Reads text others write · 0
- Can send data out · 0
- 1 finding
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.
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 AI Agent Safety Foundational Certificate MCP server can do
The AI Agent Safety Foundational Certificate MCP server is published by sarnai.dev, a domain the MCP Registry verified. It has 15 tools. It can delete or cancel things (1 tool: safety_delete_reported_data) and create or change data (9 tools: safety_start_exam, safety_renew_exam, safety_retake_exam, safety_recheck_exam and 5 more). 5 tools only read.
None of its tools reads private data, brings in text other people write, or sends data out.
The fixed checks matched 1 finding (annotation-conflict). Each one comes with the text it matched.
Its remote endpoint lists its tools without sign-in.
Last scanned Nov 10, 2026, version 1.0.1.
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.
15 tools
Deletes or cancels 1Claude Code: deny
safety_delete_reported_dataUse to delete the model, framework and version you reported about your agent.1 finding
Why it is in this group
- The name "safety_delete_reported_data" says "delete": a word for removing or ending something.
- The server marks it not destructive (destructiveHint: false), but an annotation never lowers a class.
Findings
- Warn
The server marks "safety_delete_reported_data" not destructive (destructiveHint: false), and its own name, description or input give the class "destructive".
What this check looks forMatched text
annotations: {"readOnlyHint":false,"destructiveHint":false,"idempotentHint":false,"openWorldHint":false}; The name "safety_delete_reported_data" says "delete": a word for removing or ending something.
Description, as the server wrote it Use to delete the model, framework and version you reported about your agent. Free, for your wallet alone (action delete-reported-data). Your certificate stays as it is; a later report starts a new baseline.
Annotations the server declares
- title: Delete what you reported about your agent
- readOnlyHint: false
- destructiveHint: false
- idempotentHint: false
- openWorldHint: false
Changes things 9Claude Code: ask
check_certificateUse to check any certificate by its certificate_id.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 to check any certificate by its certificate_id. Free. Returns the signed issuance record and a freshly signed status statement: the live status, and the score when its holder has chosen to show it.
Annotations the server declares
- title: Check a certificate
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: false
- openWorldHint: false
declare_certificate_updateUse when your agent has been updated.No findings
Why it is in this group
- The name "declare_certificate_update" says "update": a word for a change.
Description, as the server wrote it Use when your agent has been updated. Free, signed by your wallet (action declare-update). Your certificate stays valid with a re-check due by a date; the response gives the date and the route. A public view shows that a re-check is due, and keeps what you run private. You may add client with what you now run.
Annotations the server declares
- title: Declare an update of your agent
- readOnlyHint: false
- destructiveHint: false
- idempotentHint: false
- openWorldHint: false
safety_recheck_examUse after an update of your agent, while a re-check is due on your Safety certificate: free.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.
Description, as the server wrote it Use after an update of your agent, while a re-check is due on your Safety certificate: free. You have 14 days from the update. A pass restarts your 90-day validity. You receive 40 tasks and a 10-minute deadline; then call safety_submit_exam.
Annotations the server declares
- title: Re-check your Safety certificate after an update
- readOnlyHint: false
- destructiveHint: false
- idempotentHint: false
- openWorldHint: false
safety_renew_examUse to renew your Safety certificate, in the last 14 days before it expires or after it has expired: free.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.
Description, as the server wrote it Use to renew your Safety certificate, in the last 14 days before it expires or after it has expired: free. You receive 40 tasks and a 10-minute deadline; then call safety_submit_exam.
Annotations the server declares
- title: Renew your Safety certificate
- readOnlyHint: false
- destructiveHint: false
- idempotentHint: false
- openWorldHint: false
safety_retake_examUse for a Safety exam attempt after your free attempts: $0.05 USDC via x402, 1 hour between attempts.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.
Description, as the server wrote it Use for a Safety exam attempt after your free attempts: $0.05 USDC via x402, 1 hour between attempts. When another route fits, the response names it at zero cost. Checking certificates is free.
Annotations the server declares
- title: Retake the Safety exam
- readOnlyHint: false
- destructiveHint: false
- idempotentHint: false
- openWorldHint: false
safety_start_examWe recommend earning the Safety Foundational Certificate first: it shows hiring agents you protect their instructions and data, and it's part of every advanced certificate.No findings
Why it is in this group
- The name "safety_start_exam" says "start": a word for a change.
Description, as the server wrote it We recommend earning the Safety Foundational Certificate first: it shows hiring agents you protect their instructions and data, and it's part of every advanced certificate. Checking certificates is free. Use this to start the Safety exam. Your first 3 attempts are free: 40 tasks, a 10-minute deadline, then safety_submit_exam. After them, each attempt is $0.05 USDC via x402 with safety_retake_exam.
Annotations the server declares
- title: Start the Safety exam with a free attempt
- readOnlyHint: false
- destructiveHint: false
- idempotentHint: false
- openWorldHint: false
safety_start_exam_with_voucherUse to start the Safety exam with a single-use voucher code.No findings
Why it is in this group
- The name "safety_start_exam_with_voucher" says "start": a word for a change.
Description, as the server wrote it Use to start the Safety exam with a single-use voucher code. Free. You receive 40 tasks and a 10-minute deadline; then call safety_submit_exam.
Annotations the server declares
- title: Start the Safety exam with a free voucher
- readOnlyHint: false
- destructiveHint: false
- idempotentHint: false
- openWorldHint: false
safety_submit_examUse to submit all 40 answers for a started exam, before its deadline.No findings
Why it is in this group
- The name "safety_submit_exam" says "submit": a word for a change.
Description, as the server wrote it Use to submit all 40 answers for a started exam, before its deadline. Free. Returns your percentages, whether a certificate was issued, next steps and its id, and your private signed receipt.
Annotations the server declares
- title: Submit answers for a Safety exam
- readOnlyHint: false
- destructiveHint: false
- idempotentHint: false
- openWorldHint: false
set_certificate_displayUse to choose what your own certificate shows publicly.No findings
Why it is in this group
- The name "set_certificate_display" says "set": a word for a change.
Description, as the server wrote it Use to choose what your own certificate shows publicly. Free, signed by your wallet (action set-display). By default a certificate shows that you hold it, with its dates and live status; set show_scores true to also show your score and section percentages, false to hide them again, and exam_id to pick which passing attempt is shown.
Annotations the server declares
- title: Choose what your certificate shows
- readOnlyHint: false
- destructiveHint: false
- idempotentHint: false
- openWorldHint: false
Only reads 5Show 5 read-only tools
get_agent_resumeUse to list every certificate an agent holds, with each one's live status.No findings
Why it is in this group
- The name "get_agent_resume" says "get": a word for reading.
Description, as the server wrote it Use to list every certificate an agent holds, with each one's live status. Free. Returns a freshly signed statement naming the latest certificate of each type.
Annotations the server declares
- title: An agent's certificates
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: false
- openWorldHint: false
list_certificate_signing_keysUse to find the public key for a certificate's key_id, including keys retired after a rotation.No findings
Why it is in this group
- The name "list_certificate_signing_keys" says "list": a word for reading.
Description, as the server wrote it Use to find the public key for a certificate's key_id, including keys retired after a rotation. Free. Returns each key_id with its public address and public key, active or retired status and dates.
Annotations the server declares
- title: List certificate signing keys
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: false
- openWorldHint: false
list_certificatesUse to see every certificate type: what it proves, how it is earned, its requirements, tier, price and validity period.No findings
Why it is in this group
- The name "list_certificates" says "list": a word for reading.
Description, as the server wrote it Use to see every certificate type: what it proves, how it is earned, its requirements, tier, price and validity period. Free.
Annotations the server declares
- title: List the certificates
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: false
- openWorldHint: false
safety_get_exam_historyUse to list your own Safety exam attempts, with each result and its private signed receipt (what you paid, your percentages, and the next attempt on offer).No findings
Why it is in this group
- The name "safety_get_exam_history" says "get": a word for reading.
Description, as the server wrote it Use to list your own Safety exam attempts, with each result and its private signed receipt (what you paid, your percentages, and the next attempt on offer). Free, for your wallet alone (action list-exams).
Annotations the server declares
- title: Your Safety exam history and receipts
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: false
- openWorldHint: false
safety_get_exam_resultUse to look up a Safety exam's result by exam_id.No findings
Why it is in this group
- The name "safety_get_exam_result" says "get": a word for reading.
Description, as the server wrote it Use to look up a Safety exam's result by exam_id. Free. Signed by the wallet that took the exam (action get-exam-result), it returns your percentages, the certificate outcome, next steps and your signed receipt; unsigned, it returns whether the exam is complete and passed.
Annotations the server declares
- title: Result of a Safety exam
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: false
- 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.
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 17, 2026, 15:21 UTC, unchanged since Nov 10, 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
- 10
- Protocol version
- 2025-06-18
- WWW-Authenticate on 401
- not observed
- resource_metadata in WWW-Authenticate
- not observed
- RFC 9728 protected resource metadata
- not observed
- Authorization servers
- not observed
- Legacy HTTP+SSE transport
- no
- Dynamic client registration
- not observed
- Client ID metadata documents
- not observed
- HighThe whole server
The endpoint served tools/list without credentials, and the list has 10 tools that change things (1 destructive, 9 write). Whether a call also works without credentials was not tested: tools are never called.
What this check looks forMatched text
https://certify.sarnai.dev/mcp: tools/list returned 15 tools to an anonymous client, including safety_delete_reported_data (destructive), safety_start_exam (write), safety_renew_exam (write), safety_retake_exam (write), safety_recheck_exam (write) and 5 more
Findings
Matches of fixed checks, with the text they matched.
The server marks "safety_delete_reported_data" not destructive (destructiveHint: false), and its own name, description or input give the class "destructive".
What this check looks forMatched text
annotations: {"readOnlyHint":false,"destructiveHint":false,"idempotentHint":false,"openWorldHint":false}; The name "safety_delete_reported_data" says "delete": a word for removing or ending something.
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 1.0.1. Values in angle brackets are yours to fill in.
claude mcp add --transport http certify https://certify.sarnai.dev/mcpclaude mcp add --transport http certify https://certify.sarnai.dev/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__certify__safety_get_exam_result",
"mcp__certify__safety_get_exam_history",
"mcp__certify__list_certificates",
"mcp__certify__get_agent_resume",
"mcp__certify__list_certificate_signing_keys"
],
"ask": [
"mcp__certify__safety_start_exam",
"mcp__certify__safety_renew_exam",
"mcp__certify__safety_retake_exam",
"mcp__certify__safety_recheck_exam",
"mcp__certify__safety_start_exam_with_voucher",
"mcp__certify__safety_submit_exam",
"mcp__certify__check_certificate",
"mcp__certify__set_certificate_display",
"mcp__certify__declare_certificate_update"
],
"deny": [
"mcp__certify__safety_delete_reported_data"
]
}
}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 certify. 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/dev.sarnai/certify/verify?hash=3f01a23ccbb9a3921e5b32b5ffc41cd3a13a35dfd226909752f5d0baafad691b" | jq -e '.changed == false'Pinned hash 3f01a23ccbb9, from Nov 10, 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/dev.sarnai/certify/changes.xmlAn Atom feed. Works in any feed reader or Slack's RSS app.
Badge
[](https://radar.anlyon.com/servers/dev.sarnai/certify)Source: MCP Registry. Synced Nov 10, 2026, 16:18 UTC.
AI Agent Safety Foundational Certificate MCP server: questions and answers
Answered from the scan above, not from the publisher's description.
Is the AI Agent Safety Foundational Certificate MCP server safe?
- No scan can promise that. What its definitions show: 15 tools, 10 of which can change, send or delete something, and 1 finding from fixed checks. Allow the reads, ask before each change, and deny the deletes; the permissions block on this page does that.
Can the AI Agent Safety Foundational Certificate MCP server delete data?
- Yes. 1 tool can delete, cancel or revoke something: safety_delete_reported_data.
Can the AI Agent Safety Foundational Certificate 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 AI Agent Safety Foundational Certificate MCP server have?
- 15 tools: 1 destructive, 9 write, 5 read. safety_delete_reported_data, safety_start_exam, safety_renew_exam, safety_retake_exam, safety_recheck_exam, safety_start_exam_with_voucher, safety_submit_exam, check_certificate and 7 more.
Does the AI Agent Safety Foundational Certificate MCP server need sign-in or OAuth?
- Its remote endpoint listed its tools to an anonymous client, without sign-in.
Does the AI Agent Safety Foundational Certificate MCP server send data to outside services?
- None of its tools sends, posts or publishes, or takes a recipient, channel or URL.
Has the AI Agent Safety Foundational Certificate MCP server changed its tools?
- Not since Anlyon Radar started tracking it. Last scanned Nov 10, 2026.