Agent Zero — ERC-8004 Trust Filter MCP server
io.onchainagentintel/agent-zero-intelonchainagentintel.iostreamable-httpv1.9.0Scanned
Trust filter for ERC-8004: only ~3.8% of x402 endpoints hit 5+ payers/30d. Free MCP; x402 upgrade.
9 tools: 1 change things and 8 only read.
- Reads private data · 1
- Reads text others write · 3
- Can send data out · 2
- No findings
Tools listed without sign-in. Remote posture
With Anlyon
1 tool here changes 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 Agent Zero — ERC-8004 Trust Filter MCP server can do
The Agent Zero — ERC-8004 Trust Filter MCP server is published by onchainagentintel.io, a domain the MCP Registry verified. It has 9 tools. It can create or change data (1 tool: agent_trust_teaser), but cannot delete data. 8 tools only read.
1 tool reads private data (list_payable_x402_endpoints); 3 tools bring in text other people can write (list_payable_x402_endpoints, check_x402_endpoint and seller_check); 2 tools can send data out of your system (check_x402_endpoint and seller_check). 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 2 times since Anlyon Radar first saw them.
Last scanned Nov 18, 2026, version 1.9.0.
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
Changes things 1Claude Code: ask
Only reads 8Show 8 read-only tools
check_x402_endpointFind out if agents can actually pay you.Reads text others writeCan send data outNo findings
Why it is in this group
- The description's first verb is "find": a word for reading.
What it can reach
- 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 Find out if agents can actually pay you. Free check of an x402 seller endpoint (Seller Check). Sends only unpaid requests: confirms the route answers HTTP 402, reads the challenge (header and body, v1 or v2), checks every accepts entry (scheme, network, asset, amount, payTo, maxTimeoutSeconds, and extra.name/extra.version against the token's real EIP-712 domain), the named payment service's /supported, and /pricing, /openapi.json, /.well-known/x402 for consistency. Args: url: https URL of the paid route (required). method: "GET" or "POST" (optional; default from the seller's OpenAPI file, else GET). body: optional JSON body for POST routes (max 8 KB). Returns checked_at, overall (pass|warn|fail) and checks[] with status, reason and a fix hint. Never pays. Limits: 10 checks per caller per hour, 3 per host per hour.
get_address_recordWhat we can see about one x402 pay-to address on Base: how many registrations name it, whether they share one owner, which registration its payments are credited to, and x402 payme…No findings
Why it is in this group
- The name "get_address_record" says "get": a word for reading.
Description, as the server wrote it What we can see about one x402 pay-to address on Base: how many registrations name it, whether they share one owner, which registration its payments are credited to, and x402 payments seen (count and USDC, all time and 30 days). Facts only. Base only; only x402 settlements visible in USDC transfer logs; a payment shows money reached the address, not which service was bought. The registrations listed name the address in their own files; that is not a statement about who controls it. Paid counterpart: `GET /v1/intel/address/base/{address}` (x402, $0.01 USDC, the same price as the single-agent lookup) returns the full record with the registration list and payer breakdown. The `next_step` block carries the exact URL and price.
get_agent_summaryObjective, counts-only signals for one ERC-8004 agent.No findings
Why it is in this group
- The name "get_agent_summary" says "get": a word for reading.
Description, as the server wrote it Objective, counts-only signals for one ERC-8004 agent. Args (BOTH OPTIONAL — call with `{}` to see Agent Zero itself): chain: one of "base", "ethereum". Default: "base". agent_id: the numeric ERC-8004 agent id. Default: 19353 (Agent Zero, the agent serving this MCP endpoint). Example: `{}` returns Agent Zero on base; `{"chain": "base", "agent_id": 19353}` is the explicit form. Returns reachability, TLS validity, x402 support, payments to the agent's own address (booleans only: payments_to_agent_address, payment_attribution; payments_to_owner_wallet and payments_received are deprecated, removed 2026-12-05), MCP tool count, OpenAPI method count, well-known presence, owner-named (boolean), and a neutral status label (status_label_v2: Reachable, payment signals / Reachable, no payment signals / Stub or unreachable; the old status_label is deprecated, sunset 2026-11-07). No owner identity string and no payment amounts — those are paid. Returns an `error` field if the agent is not indexed. Use this to vet a counterparty agent before transacting. Paid counterpart: `GET /v1/intel/agent/{agent_id}?chain={chain}` (x402, USDC on base-mainnet) unlocks trust_verdict (safe_to_transact / caution / avoid / insufficient_data), readiness_score, Sybil-adjusted reputation, the reviews-from-paid-jobs (commerce-backed) breakdown, and for Transact-ready or higher agents a plain-English `brief` (due-diligence summary from our data, with sources and as-of date) — the decision-driving values this free response returns as `null` inside a `locked` map. Every response carries an `unlock` block with the exact paid endpoint, price, and payment discovery.
get_coverage_statsLive coverage of the ERC-8004 agent economy across Base and Ethereum.No findings
Why it is in this group
- The name "get_coverage_stats" says "get": a word for reading.
Description, as the server wrote it Live coverage of the ERC-8004 agent economy across Base and Ethereum. Returns counts only: how many agents are indexed, how many expose a live MCP server / OpenAPI / well-known card, total MCP tools and OpenAPI methods observed, contracts verified, EOAs and owners identified, plus a per-chain breakdown. No names, addresses, or handles. Use this to size the market and see where capability density is growing. Paid counterpart: `GET /v1/intel/market` (x402, USDC on base-mainnet) returns the full capability market map — which capabilities are saturated vs open, per-capability agent lists, and momentum. Every response from this free tool carries an `unlock` block with the exact paid endpoint, price, and payment discovery.
get_leaderboardsERC-8004 agents by ladder step, by advertised tool and method counts, and most recently indexed.No findings
Why it is in this group
- The name "get_leaderboards" says "get": a word for reading.
Description, as the server wrote it ERC-8004 agents by ladder step, by advertised tool and method counts, and most recently indexed. Ranked lists — by four-step ladder position, by MCP tool count, by OpenAPI method count (what an agent lists, not what it can do), and most-recently indexed — each with agent_id, chain, public name, a shareable detail URL, and the relevant count. Names and counts only: no owner wallets, handles, or endpoint URLs. Paid counterpart: `GET /v1/intel/trending` (x402, USDC on base-mainnet) returns trending intel — top earners, fastest-rising agents, and integration candidates ranked with the signals behind the ranking. Every response from this free tool carries an `unlock` block with the exact paid endpoint, price, and payment discovery.
list_payable_x402_endpointsDirectory of ERC-8004 agents that answered a well-formed HTTP 402 x402 challenge during our last field probe.Reads private dataReads text others writeNo findings
Why it is in this group
- The name "list_payable_x402_endpoints" says "list": a word for reading.
What it can reach
- private-dataThe description's first sentence says "directory": it reads files.
- untrusted-inputThe description's first sentence says "http": it reads web pages, which anyone can write.
Description, as the server wrote it Directory of ERC-8004 agents that answered a well-formed HTTP 402 x402 challenge during our last field probe. OBSERVATION ONLY — not a score, ranking, or endorsement. Args (ALL OPTIONAL — call with `{}` for the top 10 across both chains): chain: one of "base", "ethereum" (omit for both). Default: null. limit: 1..10 rows (free-tier cap). Default: 10. Returns per-row raw probe facts: agent_id, chain, endpoint_url, http_402_ok, network, asset, price_usd as the challenge advertised, payTo_present (boolean — the actual address is paid), sibling_agent_ids (other ERC-8004 ids served by the same gateway row — the 'shared operator' cue), probed_at. Plus `total_endpoints_valid_402` and `total_agents_valid_402` so a caller can see what fraction of the surface this free slice covers. Source: the 0946 field-test one-shot probe (unauthenticated, read-only, no payment signed). See the `field_test_url` field for the full writeup. Paid counterpart: the paid x402 /v1/intel/* endpoints return the same-origin enriched intel on any single agent in this list — owner identity, payment history, peers, trust verdict, Sybil-adjusted reputation. Every response from this free tool carries an `unlock` + `paid_upgrade` block with the exact paid endpoint, price, and payment discovery.
seller_checkFind out if agents can actually pay you.Reads text others writeCan send data outNo findings
Why it is in this group
- The description's first verb is "find": a word for reading.
What it can reach
- 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 Find out if agents can actually pay you. PAID ($0.50 USDC on Base, x402) Seller Check: an independent test of an x402 seller endpoint. Runs the free checks, then makes one real paid call to the seller (only if the seller's price is at most $0.10) and returns a report with every check, both transaction links, pass/warn/fail, findings and fix hints. One free re-run within 7 days. Payment: call once without payment to get the x402 v2 requirements (returned as `payment_required`), then repeat the call with the PAYMENT-SIGNATURE (v2) or X-PAYMENT (v1) header on this MCP HTTP request. The same service is POST https://api.onchainagentintel.io/v1/seller-check. Args: url: https URL of the paid route (required). method: "GET" or "POST" (optional). body: optional JSON body for the paid call (else the seller's own example; we never invent data). route: optional path to test on the url's host. contact_email: optional; stored, never published. public: true lets the result page be indexed (default private).
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 private-data tool breaks every one.
- Reads private data1
- Reads files, mail, messages, databases, secrets, calendars or repositories.
- Reads text others write3
- Brings in text someone else can write: web pages, inbound email, messages, issues and comments.
- Can send data out2
- Can carry data to people or systems: it sends, posts or publishes, or takes a recipient, channel or URL.
{
"permissions": {
"deny": [
"mcp__agent-zero-intel__list_payable_x402_endpoints"
]
}
}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, 19:13 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
- 1
- 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
- WarnThe whole server
The endpoint served tools/list without credentials, and the list has 1 tool that changes things (1 write). Whether a call also works without credentials was not tested: tools are never called.
What this check looks forMatched text
https://api.onchainagentintel.io/mcp: tools/list returned 9 tools to an anonymous client, including agent_trust_teaser (write)
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.
Tools1 added, 4 changed.+1 added~4 changedView diff
- Added:get_address_recordOnly reads
- Changed:agent_trust_teaserChanges thingsdescription
- Changed:get_agent_summaryOnly readsdescription
- Changed:get_leaderboardsOnly readsdescription
- Changed:get_network_teaserOnly readsdescription
Tools2 added, 3 changed.+2 added~3 changedView diff
- Added:check_x402_endpointOnly reads
- Added:seller_checkOnly reads
- Changed:agent_trust_teaserChanges thingsdescription
- Changed:get_agent_summaryOnly readsdescription
- Changed:get_network_teaserOnly readsdescription
Install
Pinned to version 1.9.0. Values in angle brackets are yours to fill in.
claude mcp add --transport http agent-zero-intel https://api.onchainagentintel.io/mcpclaude mcp add --transport http agent-zero-intel https://api.onchainagentintel.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__agent-zero-intel__get_coverage_stats",
"mcp__agent-zero-intel__get_leaderboards",
"mcp__agent-zero-intel__get_agent_summary",
"mcp__agent-zero-intel__get_address_record",
"mcp__agent-zero-intel__get_network_teaser",
"mcp__agent-zero-intel__list_payable_x402_endpoints"
],
"ask": [
"mcp__agent-zero-intel__agent_trust_teaser",
"mcp__agent-zero-intel__check_x402_endpoint",
"mcp__agent-zero-intel__seller_check"
],
"deny": []
}
}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 agent-zero-intel. Review it before you use it.
Capability tags behind 3 rules
- mcp__agent-zero-intel__list_payable_x402_endpoints private-data, untrusted-input
- mcp__agent-zero-intel__check_x402_endpoint untrusted-input, outbound
- mcp__agent-zero-intel__seller_check untrusted-input, 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.onchainagentintel/agent-zero-intel/verify?hash=df800a93b1690a6283a8cdd78d33c1d6a6f95d7fd956589ac24b78c9a3204245" | jq -e '.changed == false'Pinned hash df800a93b169, from Nov 18, 2026, 13:18 UTC.
Watch
Get told when its tools change: a new tool, a new description, a new class.
https://radar.anlyon.com/servers/io.onchainagentintel/agent-zero-intel/changes.xmlAn Atom feed. Works in any feed reader or Slack's RSS app.
Badge
[](https://radar.anlyon.com/servers/io.onchainagentintel/agent-zero-intel)Source: MCP Registry. Synced Oct 21, 2026, 18:52 UTC.
Agent Zero — ERC-8004 Trust Filter MCP server: questions and answers
Answered from the scan above, not from the publisher's description.
Is the Agent Zero — ERC-8004 Trust Filter MCP server safe?
- No scan can promise that. What its definitions show: 9 tools, 1 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 Agent Zero — ERC-8004 Trust Filter MCP server delete data?
- No tool is classed as destructive: going by their definitions, none of them deletes, cancels or revokes anything.
Can the Agent Zero — ERC-8004 Trust Filter 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 Agent Zero — ERC-8004 Trust Filter MCP server have?
- 9 tools: 1 write, 8 read. agent_trust_teaser, get_coverage_stats, get_leaderboards, get_agent_summary, get_address_record, get_network_teaser, list_payable_x402_endpoints, check_x402_endpoint and 1 more.
Does the Agent Zero — ERC-8004 Trust Filter MCP server need sign-in or OAuth?
- Its remote endpoint listed its tools to an anonymous client, without sign-in.
Does the Agent Zero — ERC-8004 Trust Filter MCP server send data to outside services?
- 2 tools can carry data out (it sends, posts or publishes, or takes a recipient, channel or URL): check_x402_endpoint and seller_check.
Has the Agent Zero — ERC-8004 Trust Filter MCP server changed its tools?
- Yes, 2 times since Anlyon Radar started tracking it, most recently on Nov 18, 2026. Every change is on this page and in its Atom feed.