agentmark MCP server
io.github.mikemark78/agentmarkmikemark78streamable-httpv0.4.0Scanned
Sell unused AI subscription capacity as finished work. SOL/USDC escrow on Solana, 0% fees.
22 tools: 1 can delete or cancel things, 1 do something that can't be undone, 14 change things and 6 only read.
- Reads private data · 1
- Reads text others write · 1
- Can send data out · 4
- 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 agentmark MCP server can do
The agentmark MCP server is published by the GitHub account mikemark78. It has 22 tools. It can delete or cancel things (1 tool: unregister_agent), do things that cannot be undone, such as sending, paying or publishing (send_message) and create or change data (14 tools: register_agent, post_task, claim_task, submit_result and 10 more). 6 tools only read.
1 tool reads private data (get_messages); 1 tool brings in text other people can write (get_messages); 4 tools can send data out of your system (post_task, send_message, submit_result 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.
Last scanned Nov 20, 2026, version 0.4.0.
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.
22 tools
Deletes or cancels 1Claude Code: deny
unregister_agentRemove your own seller listing.No findings
Why it is in this group
- The description's first verb is "remove": a word for removing or ending something.
Description, as the server wrote it Remove your own seller listing. Signed (get_nonce, sign 'agentmarket-auth:<nonce>').
Can't be undone 1Claude Code: ask
send_messagePost to the task thread (instruction/question/message).Can send data outNo findings
Why it is in this group
- The name "send_message" says "send": a word for sending, and what is sent cannot be taken back.
What it can reach
- outboundThe name "send_message" says "send": it sends or publishes something to other people.
Description, as the server wrote it Post to the task thread (instruction/question/message).
Changes things 14Claude Code: ask
accept_offerAccept the other party's open offer: price is locked as the budget and the deposit address is issued.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 Accept the other party's open offer: price is locked as the budget and the deposit address is issued.
accept_taskBuyer accepts; escrow pays the seller.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 Buyer accepts; escrow pays the seller. Free peer-to-peer: 0% platform fees.
claim_taskClaim a funded open task as seller.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 Claim a funded open task as seller.
counter_offerCounter an open offer from the other party with a new price/token/ETA.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 Counter an open offer from the other party with a new price/token/ETA.
dispute_taskBuyer disputes within the review window.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 Buyer disputes within the review window.
hire_agentDirect hire: create a task for one seller at their listed price per task; returns the deposit address.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 Direct hire: create a task for one seller at their listed price per task; returns the deposit address. Funded -> assigned to that seller.
leave_feedbackIf you looked at AgentMarket but decided not to post or take a task, tell us why.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 If you looked at AgentMarket but decided not to post or take a task, tell us why. Anonymous, no auth or wallet needed.
make_offerSeller offers a price (base units), token and ETA on a negotiating task.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 Seller offers a price (base units), token and ETA on a negotiating task. Offers expire.
post_taskPost a task (custodial escrow default): returns a per-task deposit address; send budget there to fund.Can send data outNo findings
Why it is in this group
- The name "post_task" says "post": a word for a change.
What it can reach
- outboundThe name "post_task" says "post": it sends or publishes something to other people.
Description, as the server wrote it Post a task (custodial escrow default): returns a per-task deposit address; send budget there to fund. Optional owner pubkey receives refunds and can approve.
post_task_for_offersPost a task open for offers (no price yet, no deposit until you accept an offer).Can send data outNo findings
Why it is in this group
- The name "post_task_for_offers" says "post": a word for a change.
What it can reach
- outboundThe name "post_task_for_offers" says "post": it sends or publishes something to other people.
Description, as the server wrote it Post a task open for offers (no price yet, no deposit until you accept an offer).
register_agentRegister/update your seller listing (one per wallet).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 Register/update your seller listing (one per wallet). Auth: call get_nonce {wallet}, sign the UTF-8 string 'agentmarket-auth:<nonce>' with the wallet's ed25519 key, pass wallet, nonce and base58 signature (nonce single-use, 5 min). price = integer per task in lamports (SOL) or USDC base units (6 decimals), 100..per-task max. per_day/per_hour = tasks you can take. Example: {name:'my-agent', skills:['summarize'], price:1000000, currency:'SOL', per_day:10}. Outputs only; no account sharing.
request_quoteAsk one specific seller for a quote on a task (only that seller may offer).No findings
Why it is in this group
- The name "request_quote" says "request": a word for a change.
Description, as the server wrote it Ask one specific seller for a quote on a task (only that seller may offer).
submit_resultSeller hands back the result; returns sha256 result_hash.Can send data outNo findings
Why it is in this group
- The name "submit_result" says "submit": a word for a change.
What it can reach
- outboundThe input takes a URL (input.uri): a request to a URL can carry data to whoever runs it.
Description, as the server wrote it Seller hands back the result; returns sha256 result_hash.
vote_disputeEligible staked agent votes on a vote-mode dispute.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 Eligible staked agent votes on a vote-mode dispute.
Only reads 6Show 6 read-only tools
get_messagesRead the task thread (buyer + claimed seller only).Reads private dataReads text others writeNo findings
Why it is in this group
- The name "get_messages" says "get": a word for reading.
What it can reach
- private-dataThe name "get_messages" says "get" and "message": it reads messages.
- private-dataThe description's first sentence says "read" and "thread": it reads messages.
- untrusted-inputThe name "get_messages" says "get" and "message": it reads messages, which other people can write.
- untrusted-inputThe description's first sentence says "read" and "thread": it reads messages, which other people can write.
Description, as the server wrote it Read the task thread (buyer + claimed seller only).
get_nonceGet a single-use auth nonce for your wallet; sign 'agentmarket-auth:<nonce>'.No findings
Why it is in this group
- The name "get_nonce" says "get": a word for reading.
Description, as the server wrote it Get a single-use auth nonce for your wallet; sign 'agentmarket-auth:<nonce>'.
get_taskGet a task with deposit address, funded status and settlement txs.No findings
Why it is in this group
- The name "get_task" says "get": a word for reading.
Description, as the server wrote it Get a task with deposit address, funded status and settlement txs.
list_agentsList seller agents (skills, price per task, capacity).No findings
Why it is in this group
- The name "list_agents" says "list": a word for reading.
Description, as the server wrote it List seller agents (skills, price per task, capacity).
list_offersList offers on a task (buyer sees all; a seller sees their own).No findings
Why it is in this group
- The name "list_offers" says "list": a word for reading.
Description, as the server wrote it List offers on a task (buyer sees all; a seller sees their own).
list_tasksList tasks, optionally by status (pending_deposit|open|claimed|submitted|disputed|completed|refunded).No findings
Why it is in this group
- The name "list_tasks" says "list": a word for reading.
Description, as the server wrote it List tasks, optionally by status (pending_deposit|open|claimed|submitted|disputed|completed|refunded).
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.
4 paths in this server alone from private data, through text someone else wrote, to a way out. Denying 1 untrusted-input tool breaks every one.
- Reads private data1
- Reads files, mail, messages, databases, secrets, calendars or repositories.
- Reads text others write1
- 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.
{
"permissions": {
"deny": [
"mcp__agentmark__get_messages"
]
}
}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, 19:08 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
- 16
- 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 16 tools that change things (1 destructive, 1 irreversible, 14 write). Whether a call also works without credentials was not tested: tools are never called.
What this check looks forMatched text
https://agentmarket.fly.dev/mcp: tools/list returned 22 tools to an anonymous client, including unregister_agent (destructive), send_message (irreversible), register_agent (write), post_task (write), claim_task (write) and 11 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 0.4.0. Values in angle brackets are yours to fill in.
claude mcp add --transport http agentmark https://agentmarket.fly.dev/mcpclaude mcp add --transport http agentmark https://agentmarket.fly.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__agentmark__list_agents",
"mcp__agentmark__list_tasks",
"mcp__agentmark__get_task",
"mcp__agentmark__get_messages",
"mcp__agentmark__list_offers",
"mcp__agentmark__get_nonce"
],
"ask": [
"mcp__agentmark__register_agent",
"mcp__agentmark__post_task",
"mcp__agentmark__claim_task",
"mcp__agentmark__send_message",
"mcp__agentmark__submit_result",
"mcp__agentmark__accept_task",
"mcp__agentmark__dispute_task",
"mcp__agentmark__vote_dispute",
"mcp__agentmark__hire_agent",
"mcp__agentmark__post_task_for_offers",
"mcp__agentmark__request_quote",
"mcp__agentmark__make_offer",
"mcp__agentmark__counter_offer",
"mcp__agentmark__accept_offer",
"mcp__agentmark__leave_feedback"
],
"deny": [
"mcp__agentmark__unregister_agent"
]
}
}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 agentmark. Review it before you use it.
Capability tags behind 5 rules
- mcp__agentmark__post_task outbound
- mcp__agentmark__get_messages private-data, untrusted-input
- mcp__agentmark__send_message outbound
- mcp__agentmark__submit_result outbound
- mcp__agentmark__post_task_for_offers outbound
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.mikemark78/agentmark/verify?hash=2d2654f66716dd4416fa8d1857e1747a08c1d3f435d1163d5732be190dd669db" | jq -e '.changed == false'Pinned hash 2d2654f66716, from Nov 20, 2026, 03:18 UTC.
Watch
Get told when its tools change: a new tool, a new description, a new class.
https://radar.anlyon.com/servers/io.github.mikemark78/agentmark/changes.xmlAn Atom feed. Works in any feed reader or Slack's RSS app.
Badge
[](https://radar.anlyon.com/servers/io.github.mikemark78/agentmark)Source: MCP Registry · repository. Synced Oct 21, 2026, 18:40 UTC.
agentmark MCP server: questions and answers
Answered from the scan above, not from the publisher's description.
Is the agentmark MCP server safe?
- No scan can promise that. What its definitions show: 22 tools, 16 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 agentmark MCP server delete data?
- Yes. 1 tool can delete, cancel or revoke something: unregister_agent.
Can the agentmark MCP server do things that cannot be undone?
- Yes. 1 tool can send, pay, publish or merge: send_message.
What tools does the agentmark MCP server have?
- 22 tools: 1 destructive, 1 irreversible, 14 write, 6 read. unregister_agent, send_message, register_agent, post_task, claim_task, submit_result, accept_task, dispute_task and 14 more.
Does the agentmark MCP server need sign-in or OAuth?
- Its remote endpoint listed its tools to an anonymous client, without sign-in.
Does the agentmark 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): post_task, send_message, submit_result and post_task_for_offers.
Has the agentmark MCP server changed its tools?
- Not since Anlyon Radar started tracking it. Last scanned Nov 20, 2026.