apex-edge MCP server
io.github.NonSynthetic/apex-edgeNonSyntheticstreamable-httpv1.4.0Scanned
x402 machine-payable utility edge: 19 curated MCP tools + ~450-SKU REST catalog, priced in USDC
19 tools: 1 do something that can't be undone, 15 change things and 3 only read.
- Reads private data · 1
- Reads text others write · 3
- Can send data out · 1
- 1 finding
Tools listed without sign-in. Remote posture
With Anlyon
1 tool here does something that can't be undone. 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 apex-edge MCP server can do
The apex-edge MCP server is published by the GitHub account NonSynthetic. It has 19 tools. It can do things that cannot be undone, such as sending, paying or publishing (units) and create or change data (15 tools: hash, uuid, uuid_batch, base64 and 11 more), but cannot delete data. 3 tools only read.
1 tool reads private data (corpus_search); 3 tools bring in text other people can write (price, market_pack and market_desk_brief); 1 tool can send data out of your system (units). On its own it has all three parts of a prompt-injection exfiltration path: private data, text others write, and a way 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 13, 2026, version 1.4.0.
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.
19 tools
Can't be undone 1Claude Code: ask
units$0.0020 — Unit conversion across common metric/imperial quantities.Can send data out1 finding
Why it is in this group
- The input addresses people (input.to): what it sends cannot be taken back.
- The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.
What it can reach
- outboundThe input takes a destination (input.to): what the tool writes goes there.
Findings
- Warn
The server marks "units" read-only (readOnlyHint: true), and its own name, description or input give the class "irreversible".
What this check looks forMatched text
annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":false}; The input addresses people (input.to): what it sends cannot be taken back.
Description, as the server wrote it $0.0020 — Unit conversion across common metric/imperial quantities. Use when normalizing measurements before math or display. Args: value (number), from, to (unit strings). Returns {value, from, to, result}.
Annotations the server declares
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: false
Changes things 15Claude Code: ask
base64$0.0020 — Base64 encode/decode UTF-8 text up to 32KB.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 $0.0020 — Base64 encode/decode UTF-8 text up to 32KB. Use when packing payloads for transport or decoding received blobs. Args: input (string, required), decode (boolean, default false). Returns {out}.
Annotations the server declares
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: false
board_claim$0.0020 — Claim an open board task for your agent name.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 $0.0020 — Claim an open board task for your agent name. Use when locking work found via board_list before delivering — claims are token-bound and capped 3 per task. Args: id (task id from board_list, required), claimer, payout_wallet (0x Base EOA, required for usdc tasks). Returns claim token + task detail.
Annotations the server declares
- readOnlyHint: false
- destructiveHint: false
- idempotentHint: false
- openWorldHint: false
checklist$0.0500 — Static 32-item smart-contract/token security checklist — deterministic gates, no LLM.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 $0.0500 — Static 32-item smart-contract/token security checklist — deterministic gates, no LLM. Use when running a pre-buy risk screen or release preflight. No args. Returns the full checklist JSON.
Annotations the server declares
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: false
hash$0.0020 — SHA/MD5 digest of UTF-8 text.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 $0.0020 — SHA/MD5 digest of UTF-8 text. Use when you need a checksum, content address or dedupe key — not for password hashing. Args: input (string, required), algo (sha256|sha384|sha512|sha1|md5, default sha256). Returns {algo, hex}. Pure function, no state.
Annotations the server declares
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: false
intel_board_economics$2.0000 — Verified dossier: agent-money board economics across 18 venues — payout states, listing counts, claim/settle evidence from our own ledgers.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 $2.0000 — Verified dossier: agent-money board economics across 18 venues — payout states, listing counts, claim/settle evidence from our own ledgers. Use when evaluating which agent-work venue actually pays. No args. Returns the dossier JSON.
Annotations the server declares
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: false
jwt_decode$0.0020 — Decode a JWT into header+payload JSON WITHOUT signature verification.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 $0.0020 — Decode a JWT into header+payload JSON WITHOUT signature verification. Use when inspecting a token's claims or expiry — never for auth decisions. Arg: token (string, required). Returns {verified:false, header, payload}.
Annotations the server declares
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: false
market_desk_brief$0.0500 — Compiled market desk brief — funding extremes, basis moves, notable shifts as markdown.Reads text others writeNo 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.
Description, as the server wrote it $0.0500 — Compiled market desk brief — funding extremes, basis moves, notable shifts as markdown. Use when you want a human-readable market digest or a report attachment. Arg: asset (string, optional, default BTC). Returns text/markdown.
Annotations the server declares
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: true
market_pack$0.0050 — Bundled crypto market snapshot — spot, funding, basis in one call (cheaper than the singles).Reads text others writeNo 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.
Description, as the server wrote it $0.0050 — Bundled crypto market snapshot — spot, funding, basis in one call (cheaper than the singles). Use when you need a quick market read before sizing a position or quoting a rate. Arg: asset (string, optional, default BTC). Returns the pack JSON.
Annotations the server declares
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: true
price$0.0020 — Spot price from Coinbase's public API (no key needed; verifiable against the same public feed).Reads text others writeNo 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.
Description, as the server wrote it $0.0020 — Spot price from Coinbase's public API (no key needed; verifiable against the same public feed). Use when you need a quick 'price of X' check or a rate for payment math. Arg: pair (e.g. BTC-USD, default BTC-USD). Returns {pair, spot, ts}. Cached 30s.
Annotations the server declares
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: true
semver$0.0020 — Numeric-core semver compare of two versions.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 $0.0020 — Numeric-core semver compare of two versions. Use when deciding upgrade/downgrade ordering — pre-release metadata ignored, so not for prerelease channels. Args: a, b (strings, required). Returns {cmp: -1|0|1}.
Annotations the server declares
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: false
slugify$0.0020 — URL-safe slug from arbitrary text (NFKD, ascii, dash-joined).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 $0.0020 — URL-safe slug from arbitrary text (NFKD, ascii, dash-joined). Use when generating permalinks or machine ids from titles or names. Arg: input (string, required). Returns {slug}.
Annotations the server declares
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: false
time$0.0020 — Current UTC timestamp.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 $0.0020 — Current UTC timestamp. Use when syncing clocks, computing deadlines or stamping data. No args. Returns {iso, unix_ms} — ISO-8601 string plus milliseconds since epoch.
Annotations the server declares
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: false
uuid$0.0020 — One random UUID v4 (RFC 4122).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 $0.0020 — One random UUID v4 (RFC 4122). Use when you need a request id, dedupe key or fixture value — for bulk ids use uuid_batch instead. No args. Returns {uuid}.
Annotations the server declares
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: false
uuid_batch$0.0020 — 1-1000 random UUID v4s in one call.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 $0.0020 — 1-1000 random UUID v4s in one call. Use when seeding fixtures or idempotency pools — cheaper per-id than repeated uuid calls. Arg: count (number, default 10, max 1000). Returns {count, uuids[]}.
Annotations the server declares
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: false
wordcount$0.0020 — Text statistics for a document.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 $0.0020 — Text statistics for a document. Use when checking length limits, estimating token budget or comparing docs. Arg: input (string, required, ≤64KB). Returns {chars, words, lines, sentences}.
Annotations the server declares
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: false
Only reads 3Show 3 read-only tools
board_list$0.0020 — Open tasks on the apex board — claimable work for agents (earn loop: claim → deliver → cr credit → spend on tools).No findings
Why it is in this group
- The name "board_list" says "list": a word for reading.
Description, as the server wrote it $0.0020 — Open tasks on the apex board — claimable work for agents (earn loop: claim → deliver → cr credit → spend on tools). Use when hunting paid work your agent can claim — free entry, no wallet needed. Args: status, q, min_reward, limit (all optional). Returns task list JSON.
Annotations the server declares
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: false
corpus_search$0.0100 — Keyword search over a verified research corpus — agent-money venues, x402 economics, settlement evidence; every doc probed live and dated.Reads private dataNo findings
Why it is in this group
- The name "corpus_search" says "search": a word for reading.
What it can reach
- private-dataThe description's first sentence says "doc": it reads documents.
Description, as the server wrote it $0.0100 — Keyword search over a verified research corpus — agent-money venues, x402 economics, settlement evidence; every doc probed live and dated. Use when you need cited facts about agent payments/markets instead of guessing. Arg: q (string, required). Returns {docs_indexed, hits, results[{path,title,snippet,score}]}.
Annotations the server declares
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: false
json_validate$0.0020 — JSON syntax validity check.No findings
Why it is in this group
- The name "json_validate" says "validate": a word for reading.
Description, as the server wrote it $0.0020 — JSON syntax validity check. Use when gating a blob before parsing, storing or forwarding it. Arg: input (string, required, ≤64KB). Returns {valid, error?}.
Annotations the server declares
- 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.
3 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 write3
- 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__apex-edge__units"
]
}
}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, 18:21 UTC, unchanged since Nov 20, 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
- 12
- 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 12 tools that change things (12 write). Whether a call also works without credentials was not tested: tools are never called.
What this check looks forMatched text
https://api.nonsynthetic.net/mcp: tools/list returned 14 tools to an anonymous client, including board_claim (write), checklist (write), market_pack (write), market_desk_brief (write), checkpoint_save (write) and 7 more
Posture changes
- Nov 20, 2026
- Tools that change things, listed without credentials: 16 to 12
Findings
Matches of fixed checks, with the text they matched.
- Warnunits
The server marks "units" read-only (readOnlyHint: true), and its own name, description or input give the class "irreversible".
What this check looks forMatched text
annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":false}; The input addresses people (input.to): what it sends cannot be taken back.
History
Changes to its tool definitions and to its remote endpoint, newest first.
PostureRemote posture changed. Tools that change things, listed without credentials: 16 -> 12.View diff
- Tools that change things, listed without credentials: 16 to 12
Install
Pinned to version 1.4.0. Values in angle brackets are yours to fill in.
claude mcp add --transport http apex-edge https://api.nonsynthetic.net/mcpclaude mcp add --transport http apex-edge https://api.nonsynthetic.net/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__apex-edge__json_validate",
"mcp__apex-edge__corpus_search",
"mcp__apex-edge__board_list"
],
"ask": [
"mcp__apex-edge__hash",
"mcp__apex-edge__uuid",
"mcp__apex-edge__uuid_batch",
"mcp__apex-edge__base64",
"mcp__apex-edge__time",
"mcp__apex-edge__price",
"mcp__apex-edge__jwt_decode",
"mcp__apex-edge__wordcount",
"mcp__apex-edge__slugify",
"mcp__apex-edge__semver",
"mcp__apex-edge__units",
"mcp__apex-edge__board_claim",
"mcp__apex-edge__checklist",
"mcp__apex-edge__market_pack",
"mcp__apex-edge__market_desk_brief",
"mcp__apex-edge__intel_board_economics"
],
"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 apex-edge. Review it before you use it.
Capability tags behind 5 rules
- mcp__apex-edge__price untrusted-input
- mcp__apex-edge__units outbound
- mcp__apex-edge__corpus_search private-data
- mcp__apex-edge__market_pack untrusted-input
- mcp__apex-edge__market_desk_brief 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/io.github.NonSynthetic/apex-edge/verify?hash=2b4f9655add149f8ffc647bc359882e3018b88e7acaee9bddd8d027f0702d3bf" | jq -e '.changed == false'Pinned hash 2b4f9655add1, from Nov 13, 2026, 05:18 UTC.
Watch
Get told when its tools change: a new tool, a new description, a new class.
https://radar.anlyon.com/servers/io.github.NonSynthetic/apex-edge/changes.xmlAn Atom feed. Works in any feed reader or Slack's RSS app.
Badge
[](https://radar.anlyon.com/servers/io.github.NonSynthetic/apex-edge)Source: MCP Registry. Synced Nov 13, 2026, 05:17 UTC.
apex-edge MCP server: questions and answers
Answered from the scan above, not from the publisher's description.
Is the apex-edge MCP server safe?
- No scan can promise that. What its definitions show: 19 tools, 16 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 apex-edge MCP server delete data?
- No tool is classed as destructive: going by their definitions, none of them deletes, cancels or revokes anything.
Can the apex-edge MCP server do things that cannot be undone?
- Yes. 1 tool can send, pay, publish or merge: units.
What tools does the apex-edge MCP server have?
- 19 tools: 1 irreversible, 15 write, 3 read. units, hash, uuid, uuid_batch, base64, time, price, jwt_decode and 11 more.
Does the apex-edge MCP server need sign-in or OAuth?
- Its remote endpoint listed its tools to an anonymous client, without sign-in.
Does the apex-edge 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): units.
Has the apex-edge MCP server changed its tools?
- Not since Anlyon Radar started tracking it. Last scanned Nov 13, 2026.