Skip to content

SpendPreflight MCP server

io.github.spendpreflight/spendpreflightspendpreflightstreamable-httpv1.2.0-metadata.9Scanned

Preflight x402 payments. Base ERC-8004 agent #1151628; wallet binding verified Nov 3, 2026.

17 tools: 2 do something that can't be undone, 13 change things and 2 only read.

Tools: 0 delete, 2 irreversible, 13 write, 2 read.
  • Reads private data · 0
  • Reads text others write · 11
  • Can send data out · 5
  • 28 findings

Tools listed without sign-in. Remote posture

With Anlyon

2 tools here do 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.
Govern this server →

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 SpendPreflight MCP server can do

The SpendPreflight MCP server is published by the GitHub account spendpreflight. It has 17 tools. It can do things that cannot be undone, such as sending, paying or publishing (preflight_payment and payment_receipt) and create or change data (13 tools: check_payee, lint_endpoint, check_token, verify_payee_binding and 9 more), but cannot delete data. 2 tools only read.

11 tools bring in text other people can write (preflight_payment, check_payee, lint_endpoint and 8 more); 5 tools can send data out of your system (lint_endpoint, quote_check, listing_guard and 2 more).

The fixed checks matched 28 findings (base64-blob, long-description, annotation-conflict and credentials-mention). Each one comes with the text it matched.

Its remote endpoint lists its tools without sign-in, and its tools changed 8 times since Anlyon Radar first saw them.

Last scanned Nov 15, 2026, version 1.2.0-metadata.9.

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.

17 tools

Can't be undone 2Claude Code: ask
  • payment_receiptUse when you need a signed observation of an existing Base USDC transfer.Reads text others write2 findings

    Why it is in this group

    • The name "payment_receipt" says "payment": a word that usually means money moves, which cannot be taken back.
    • 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.

    Findings

    • Warn

      The input schema at properties.pay_to.description contains a 42-character base64 string.

      Matched text

      inputSchema.properties.pay_to.description: "...ss string. Example: 0xBf4164e07552e4c1b1B1597E72C7acAA3F0a1c58. Required with amou..."

      What this check looks for
    • Warn

      The server marks "payment_receipt" read-only (readOnlyHint: true), and its own name, description or input give the class "irreversible".

      Matched text

      annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":false,"openWorldHint":true}; The name "payment_receipt" says "payment": a word that usually means money moves, which cannot be taken back.

      What this check looks for
    Description, as the server wrote it
    Use when you need a signed observation of an existing Base USDC transfer. Verify a native Base USDC transfer by tx_hash (optional log_index), or exact pay_to + decimal amount + past start/end (max 90 days). Ambiguous matches fail without charge. Signed observation, OFAC flags, confirmations and printable HTML. Not invoice ownership or finality proof. Price $0.01 USDC on Base. No free trial. Price: 0.01 USDC. Minimal input example: {"tx_hash":"0x9ede68ce1d33fd5d49d402d11de3127710102235ac5bdf020344345dfeb8ce4f"}. Illustrative output excerpt: {"result":{"kind":"payment_receipt","receipt":{"amount":"0.050000","time":"2026-10-22T22:07:07.000Z"}},"jws":"illustrative-not-a-live-signature"}. Illustrative output excerpt, not a live result, signature, approval or guarantee. Freshness/data/availability checks still apply. How to pay (MCP): No shared free trial for this tool. Use an x402-capable MCP client. A generic MCP connector without payment support cannot complete a paid continuation. On payment required, STOP. Inspect the exact accepts amount, asset, network and payTo; obtain explicit wallet/budget approval before continuing. Only after approval, let the configured x402 MCP client attach its signed payload in _meta["x402/payment"] and retry the same tools/call arguments. Never send a separate USDC transfer or paste keys into tool arguments. Never silently fall back from free access to payment. If a signed attempt has an unknown outcome, inspect that authorization/transaction; do not create another automatically.

    Annotations the server declares

    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: false
    • openWorldHint: true
  • preflight_paymentUse when you are about to sign an x402 payment for allow, hold or block, reasons and a signed decision receipt.Reads text others write5 findings

    Why it is in this group

    • The name "preflight_payment" says "payment": a word that usually means money moves, which cannot be taken back.
    • 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.

    Findings

    • Warn

      The input schema at properties.challenge.description contains a 42-character base64 string.

      Matched text

      inputSchema.properties.challenge.description: "...4bda02913","payTo":"0xBf4164e07552e4c1b1B1597E72C7acAA3F0a1c58","amount":"10000","..."

      What this check looks for
    • Warn

      The input schema at properties.cart.properties.pay_to.description contains a 42-character base64 string.

      Matched text

      inputSchema.properties.cart.properties.pay_to.description: "...ss string. Example: 0xBf4164e07552e4c1b1B1597E72C7acAA3F0a1c58."

      What this check looks for
    • Warn

      The input schema at properties.cart.description contains a 42-character base64 string.

      Matched text

      inputSchema.properties.cart.description: "...usd":1.25,"pay_to":"0xBf4164e07552e4c1b1B1597E72C7acAA3F0a1c58"}. total_usd is a U..."

      What this check looks for
    • Info

      The description is 3,123 characters long, over the 2,000 this check looks for.

      Matched text

      description ends: "...ome, inspect that authorization/transaction; do not create another automatically."

      What this check looks for
    • Warn

      The server marks "preflight_payment" read-only (readOnlyHint: true), and its own name, description or input give the class "irreversible".

      Matched text

      annotations: {"readOnlyHint":true,"idempotentHint":false,"openWorldHint":true}; The name "preflight_payment" says "payment": a word that usually means money moves, which cannot be taken back.

      What this check looks for
    Description, as the server wrote it
    Use when you are about to sign an x402 payment for allow, hold or block, reasons and a signed decision receipt. Covers payee sanctions screening, spending rules and cached price/recipient evidence where available. For separate evidence, use lint_endpoint/listing_guard for technical seller-metadata checks, verify_payee_binding for ERC-8478 registry-to-wallet binding, and payment_receipt for a native Base USDC transfer observation after payment. These tools have separate scopes and may have separate charges; Preflight does not execute them automatically or verify delivered API content. A transfer receipt is not delivery proof, wallet binding is not seller-safety proof, and an allow decision does not grant payment authority. Call this before any x402/USDC payment to evaluate the actual challenge or cart before signing. Spend preflight for AI agents. Send the x402 PaymentRequired body (v1 or v2) you received, or a cart, plus optional rules (per-payment max, hold threshold, daily cap, allowlists, new-domain days, endpoint category price multiple, liveness and payTo changes). Returns allow|hold|block with reasons, the cheapest acceptable payment option, and a receipt id. Read-only: never signs, settles, or holds funds. Use before every payment. It includes payee screening in its spending-rule decision. For standalone wallet, name, or domain screening without a payment, use check_payee instead. Each repeat call creates a new receipt and may consume another trial or paid call. Price $0.02 per call (free trial first). Price: 0.02 USDC. Minimal input example: {"cart":{"merchant_name":"Example Shop LLC","total_usd":0.01},"rules":{"max_per_payment_usd":1},"spent_today_usd":0}. Illustrative output excerpt: {"decision":"hold","reasons":["hold: missing fresh index evidence"],"chosen_option":null,"options_evaluated":0,"payee":null,"index_evidence":{"status":"unknown","observed_at":null,"note":"Missing evidence remains unknown."},"receipt":{"id":"illustration","ts":"2026-10-30T00:00:00Z","input_sha256":"0000000000000000000000000000000000000000000000000000000000000000","decision":"hold","jws":"illustrative-not-a-live-signature"}}. Illustrative output excerpt, not a live result, signature, approval or guarantee. Freshness/data/availability checks still apply. How to pay (MCP): First use the existing shared allowance of 3 eligible calls per IP per UTC day, without payment metadata. This is shared with eligible HTTP trials, not a new allowance. Use an x402-capable MCP client. A generic MCP connector without payment support cannot complete a paid continuation. On payment required, STOP. Inspect the exact accepts amount, asset, network and payTo; obtain explicit wallet/budget approval before continuing. Only after approval, let the configured x402 MCP client attach its signed payload in _meta["x402/payment"] and retry the same tools/call arguments. Never send a separate USDC transfer or paste keys into tool arguments. Never silently fall back from free access to payment. If a signed attempt has an unknown outcome, inspect that authorization/transaction; do not create another automatically.

    Annotations the server declares

    • readOnlyHint: true
    • openWorldHint: true
    • idempotentHint: false
Changes things 13Claude Code: ask
  • check_endpointUse when you need cached public endpoint evidence.Can send data out2 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

    • outboundThe input takes a URL (input.url): a request to a URL can carry data to whoever runs it.

    Findings

    • Info

      The input schema at properties.url.description mentions "credentials".

      Matched text

      inputSchema.properties.url.description: "...resource URL string, including path; no credentials. Example: https://api.example.com/data."

      What this check looks for
    • Info

      The description is 2,021 characters long, over the 2,000 this check looks for.

      Matched text

      description ends: "...ome, inspect that authorization/transaction; do not create another automatically."

      What this check looks for
    Description, as the server wrote it
    Use when you need cached public endpoint evidence. Cached x402 endpoint observations, operated by SpendPreflight. Returns a transparent score (or unknown), its components and seven-day unpaid probe history for an already publicly listed URL. Never fetches the caller's URL or pays the endpoint. Unlisted, unavailable or stale data is not charged. This is not a paid-delivery or compliance guarantee. Provide the exact resource URL; add method when the URL has multiple advertised methods. Use for endpoint history and price context before paying. Use check_payee for wallet/name screening or preflight_payment for an allow/hold/block decision. Price $0.005 per cached lookup (shared free trial first). An unknown score means missing evidence; repeated paid calls may incur another charge. Price: 0.005 USDC. Minimal input example: {"url":"https://api.spendpreflight.com/v1/check","method":"GET"}. Illustrative output excerpt: {"score":null,"components":{},"history":[],"disclaimer":"An unknown score is missing evidence, not an endpoint failure."}. Illustrative output excerpt, not a live result, signature, approval or guarantee. Freshness/data/availability checks still apply. How to pay (MCP): First use the existing shared allowance of 3 eligible calls per IP per UTC day, without payment metadata. This is shared with eligible HTTP trials, not a new allowance. Use an x402-capable MCP client. A generic MCP connector without payment support cannot complete a paid continuation. On payment required, STOP. Inspect the exact accepts amount, asset, network and payTo; obtain explicit wallet/budget approval before continuing. Only after approval, let the configured x402 MCP client attach its signed payload in _meta["x402/payment"] and retry the same tools/call arguments. Never send a separate USDC transfer or paste keys into tool arguments. Never silently fall back from free access to payment. If a signed attempt has an unknown outcome, inspect that authorization/transaction; do not create another automatically.

    Annotations the server declares

    • readOnlyHint: true
    • openWorldHint: false
    • idempotentHint: true
  • check_listingUse when an existing integration calls the listing_guard compatibility alias.Reads text others writeCan send data out2 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.
    • outboundThe input takes a URL (input.url): a request to a URL can carry data to whoever runs it.

    Findings

    • Info

      The input schema at properties.url.description mentions "credentials".

      Matched text

      inputSchema.properties.url.description: "...resource URL string, including path; no credentials. Example: https://api.example.com/data."

      What this check looks for
    • Info

      The description is 2,076 characters long, over the 2,000 this check looks for.

      Matched text

      description ends: "...ome, inspect that authorization/transaction; do not create another automatically."

      What this check looks for
    Description, as the server wrote it
    Use when an existing integration calls the listing_guard compatibility alias. Compatibility alias of listing_guard for existing clients. Prefer listing_guard for a new detailed metadata check; these are the same operation, not two independent checks. Bazaar Listing Guard: CDP technical validation plus a bounded heuristic metadata screen for agent-directed instructions, embedded URLs, tool calls, invisible Unicode, long fields and prices above 100x a fresh Scout category median. Returns riskScore, flagged spans and rewrite suggestions. Not Coinbase policy certification or guaranteed detection. Public HTTPS only; no target payment or request bodies stored. Price $0.05 USDC on Base; shared trial may apply. Price: 0.05 USDC. Minimal input example: {"url":"https://api.spendpreflight.com/v1/check","method":"GET"}. Illustrative output excerpt: {"cdpPreflight":{"valid":true,"preflight":[],"index":null},"riskScore":0,"flags":[],"suggestedRewrite":[],"screenComplete":true,"priceCompared":false,"indexActive":null,"example_note":"Illustrative response only; not a live result or safety guarantee."}. Illustrative output excerpt, not a live result, signature, approval or guarantee. Freshness/data/availability checks still apply. How to pay (MCP): First use the existing shared allowance of 3 eligible calls per IP per UTC day, without payment metadata. This is shared with eligible HTTP trials, not a new allowance. Use an x402-capable MCP client. A generic MCP connector without payment support cannot complete a paid continuation. On payment required, STOP. Inspect the exact accepts amount, asset, network and payTo; obtain explicit wallet/budget approval before continuing. Only after approval, let the configured x402 MCP client attach its signed payload in _meta["x402/payment"] and retry the same tools/call arguments. Never send a separate USDC transfer or paste keys into tool arguments. Never silently fall back from free access to payment. If a signed attempt has an unknown outcome, inspect that authorization/transaction; do not create another automatically.

    Annotations the server declares

    • readOnlyHint: true
    • openWorldHint: true
  • check_payeeUse when you need standalone payee screening without a payment decision.Reads text others write1 finding

    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.

    Findings

    • Warn

      The input schema at properties.address.description contains a 42-character base64 string.

      Matched text

      inputSchema.properties.address.description: "...Base-only. Example: 0xBf4164e07552e4c1b1B1597E72C7acAA3F0a1c58. At least one of do..."

      What this check looks for
    Description, as the server wrote it
    Use when you need standalone payee screening without a payment decision. Payee risk check. Screens a payee wallet address and name against the OFAC SDN sanctions list and checks the merchant domain's registration age and DNS. Returns risk low|medium|high|unknown with flags such as sanctioned_address, sanctioned_name_possible, new_domain. Call this before a USDC payment when you only have a payee name, wallet or domain. Provide at least one of domain, address, or name; inputs may be combined, and domain accepts a hostname or URL. For an x402 challenge or cart and an allow/hold/block spending-rule decision, use preflight_payment instead. Price $0.01 per call (free trial first). Price: 0.01 USDC. Minimal input example: {"name":"Example Shop LLC"}. Illustrative output excerpt: {"risk":"low","flags":[],"sanctions":{"address_match":null,"name_matches":[]},"domain":null,"data_as_of":"2026-10-30T00:00:00Z","disclaimer":"Illustrative non-match, not sanctions clearance."}. Illustrative output excerpt, not a live result, signature, approval or guarantee. Freshness/data/availability checks still apply. How to pay (MCP): First use the existing shared allowance of 3 eligible calls per IP per UTC day, without payment metadata. This is shared with eligible HTTP trials, not a new allowance. Use an x402-capable MCP client. A generic MCP connector without payment support cannot complete a paid continuation. On payment required, STOP. Inspect the exact accepts amount, asset, network and payTo; obtain explicit wallet/budget approval before continuing. Only after approval, let the configured x402 MCP client attach its signed payload in _meta["x402/payment"] and retry the same tools/call arguments. Never send a separate USDC transfer or paste keys into tool arguments. Never silently fall back from free access to payment. If a signed attempt has an unknown outcome, inspect that authorization/transaction; do not create another automatically.

    Annotations the server declares

    • readOnlyHint: true
    • openWorldHint: true
    • idempotentHint: true
  • check_tokenUse when you need a bounded Base token assessment with explicit unassessed areas.Reads text others write4 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.

    Findings

    • Warn

      The description contains a 42-character base64 string.

      Matched text

      description: "...xample: {"address":"0x833589fCD6eDb6E08f4c7C32D4f71b54bdA02913","chain":"eip155:84..."

      What this check looks for
    • Warn

      The input schema at properties.address.description contains a 42-character base64 string.

      Matched text

      inputSchema.properties.address.description: "...ss string. Example: 0xBf4164e07552e4c1b1B1597E72C7acAA3F0a1c58."

      What this check looks for
    • Warn

      The input schema at description contains a 42-character base64 string.

      Matched text

      inputSchema.description: "...uments: {"address":"0x833589fCD6eDb6E08f4c7C32D4f71b54bdA02913","chain":"eip155:84..."

      What this check looks for
    • Info

      The description is 2,118 characters long, over the 2,000 this check looks for.

      Matched text

      description ends: "...ome, inspect that authorization/transaction; do not create another automatically."

      What this check looks for
    Description, as the server wrote it
    Use when you need a bounded Base token assessment with explicit unassessed areas. Base contract safety signals from bounded public-chain reads and OFAC. Basic: deployed bytecode and SDN match. Deep adds standard owner(), paused(), totalSupply(), decimals() and ERC-1967 storage observations, exact ERC-1167 minimal-proxy target detection, a Sourcify runtime-verification claim matched to current bytecode, and direct-deployment age when independently supported by the creation receipt. Honeypot/sellability, liquidity locks, holder concentration and scam lists are NOT assessed. Missing data is unknown; never a safe-to-buy verdict. Informational, not financial advice. Basic $0.02; deep $0.05. Shared trial first. Price: 0.02 basic; 0.05 deep USDC. Minimal input example: {"address":"0x833589fCD6eDb6E08f4c7C32D4f71b54bdA02913","chain":"eip155:8453","tier":"basic"}. Illustrative output excerpt: {"verdict":"unknown","assessment_complete":false,"unassessed":["honeypot_and_sellability","mint_and_blacklist_privileges","liquidity_lock_or_removability","holder_concentration","non_OFAC_scam_lists","complete_owner_privilege_and_proxy_analysis"]}. Illustrative output excerpt, not a live result, signature, approval or guarantee. Freshness/data/availability checks still apply. How to pay (MCP): First use the existing shared allowance of 3 eligible calls per IP per UTC day, without payment metadata. This is shared with eligible HTTP trials, not a new allowance. Use an x402-capable MCP client. A generic MCP connector without payment support cannot complete a paid continuation. On payment required, STOP. Inspect the exact accepts amount, asset, network and payTo; obtain explicit wallet/budget approval before continuing. Only after approval, let the configured x402 MCP client attach its signed payload in _meta["x402/payment"] and retry the same tools/call arguments. Never send a separate USDC transfer or paste keys into tool arguments. Never silently fall back from free access to payment. If a signed attempt has an unknown outcome, inspect that authorization/transaction; do not create another automatically.

    Annotations the server declares

    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: false
    • openWorldHint: true
  • domain_authUse when you need public DNS email-authentication observations for a domain.Reads text others write1 finding

    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.

    Findings

    • Info

      The input schema at properties.domain.description mentions "credentials".

      Matched text

      inputSchema.properties.domain.description: "...ing without scheme, path, IP address or credentials. Example: example.com."

      What this check looks for
    Description, as the server wrote it
    Use when you need public DNS email-authentication observations for a domain. Observe a public domain’s MX, SPF, direct DMARC record, supplied DKIM selectors and resolver-reported DNSSEC evidence.0.005 USDC per usable report, no shared trial. Fixed public Cloudflare DoH only, at most7 queries/shared8s/128KiB, no SMTP, sender authentication test, DNS changes or query storage. Missing selectors and inherited DMARC policy stay unknown. Published records and resolver AD are not proof of inbox delivery, alignment, DNSSEC validation by this service or a secure domain. Price: 0.005 USDC. Minimal input example: {"domain":"example.org","selectors":["selector1"]}. Illustrative output excerpt: {"spf":{"status":"unknown"},"dmarc":{"status":"unknown","alignment_verified":false},"mx":{"status":"unknown","smtp_reachability":"untested"}}. Illustrative output excerpt, not a live result, signature, approval or guarantee. Freshness/data/availability checks still apply. How to pay (MCP): No shared free trial for this tool. Use an x402-capable MCP client. A generic MCP connector without payment support cannot complete a paid continuation. On payment required, STOP. Inspect the exact accepts amount, asset, network and payTo; obtain explicit wallet/budget approval before continuing. Only after approval, let the configured x402 MCP client attach its signed payload in _meta["x402/payment"] and retry the same tools/call arguments. Never send a separate USDC transfer or paste keys into tool arguments. Never silently fall back from free access to payment. If a signed attempt has an unknown outcome, inspect that authorization/transaction; do not create another automatically.

    Annotations the server declares

    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: false
    • openWorldHint: true
  • lint_endpointUse when you sell via x402 and want a free check of your public payment challenge.Reads text others writeCan send data out2 findings

    Why it is in this group

    • Nothing in its name or description says what it does, so it is counted as a write.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    • outboundThe input takes a URL (input.url): a request to a URL can carry data to whoever runs it.

    Findings

    • Info

      The description mentions "credentials".

      Matched text

      description: "...ar acceptance guarantees. No payment or credentials; GET cannot validate a POST-only endpoi..."

      What this check looks for
    • Info

      The input schema at properties.url.description mentions "credentials".

      Matched text

      inputSchema.properties.url.description: "...resource URL string, including path; no credentials. Example: https://api.example.com/data."

      What this check looks for
    Description, as the server wrote it
    Use when you sell via x402 and want a free check of your public payment challenge. Free x402 Lint by SpendPreflight. Includes a free passive Security Check summary; optional full report with fixes is $0.10, no automatic payment. Checks a public HTTPS endpoint with an unpaid GET: v1/v2 challenge shape, payTo/asset/network syntax, price, Bazaar metadata and browser CORS headers. Returns a heuristic Listing Guard riskScore (0–100 or unknown), fixes and a public seven-day report link. Positive scores include optional paid Listing Guard and monitoring links; no paid call is made. Scores are not safety or Bazaar acceptance guarantees. No payment or credentials; GET cannot validate a POST-only endpoint or paid delivery. Do not submit private URLs. CI Action: https://github.com/spendpreflight/x402-lint-action; report badge: /v1/lint/reports/{id}/badge.svg reads only the existing seven-day snapshot, never runs a check.

    Annotations the server declares

    • readOnlyHint: false
    • destructiveHint: false
    • idempotentHint: false
    • openWorldHint: true
  • listing_guardUse when you need a detailed public listing metadata check.Reads text others writeCan send data out1 finding

    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.
    • outboundThe input takes a URL (input.url): a request to a URL can carry data to whoever runs it.

    Findings

    • Info

      The input schema at properties.url.description mentions "credentials".

      Matched text

      inputSchema.properties.url.description: "...resource URL string, including path; no credentials. Example: https://api.example.com/data."

      What this check looks for
    Description, as the server wrote it
    Use when you need a detailed public listing metadata check. Bazaar Listing Guard: CDP technical validation plus a bounded heuristic metadata screen for agent-directed instructions, embedded URLs, tool calls, invisible Unicode, long fields and prices above 100x a fresh Scout category median. Returns riskScore, flagged spans and rewrite suggestions. Not Coinbase policy certification or guaranteed detection. Public HTTPS only; no target payment or request bodies stored. Price $0.05 USDC on Base; shared trial may apply. Price: 0.05 USDC. Minimal input example: {"url":"https://api.spendpreflight.com/v1/check","method":"GET"}. Illustrative output excerpt: {"cdpPreflight":{"valid":true,"preflight":[],"index":null},"riskScore":0,"flags":[],"suggestedRewrite":[],"screenComplete":true,"priceCompared":false,"indexActive":null,"example_note":"Illustrative response only; not a live result or safety guarantee."}. Illustrative output excerpt, not a live result, signature, approval or guarantee. Freshness/data/availability checks still apply. How to pay (MCP): First use the existing shared allowance of 3 eligible calls per IP per UTC day, without payment metadata. This is shared with eligible HTTP trials, not a new allowance. Use an x402-capable MCP client. A generic MCP connector without payment support cannot complete a paid continuation. On payment required, STOP. Inspect the exact accepts amount, asset, network and payTo; obtain explicit wallet/budget approval before continuing. Only after approval, let the configured x402 MCP client attach its signed payload in _meta["x402/payment"] and retry the same tools/call arguments. Never send a separate USDC transfer or paste keys into tool arguments. Never silently fall back from free access to payment. If a signed attempt has an unknown outcome, inspect that authorization/transaction; do not create another automatically.

    Annotations the server declares

    • readOnlyHint: true
    • openWorldHint: true
  • quote_checkUse when you need an exact quote compared with cached public price and recipient evidence.Can send data out2 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

    • outboundThe input takes a URL (input.url): a request to a URL can carry data to whoever runs it.

    Findings

    • Info

      The input schema at properties.url.description mentions "credentials".

      Matched text

      inputSchema.properties.url.description: "...resource URL string, including path; no credentials. Use the current published example in O..."

      What this check looks for
    • Warn

      The input schema at properties.payTo.description contains a 42-character base64 string.

      Matched text

      inputSchema.properties.payTo.description: "...e address. Example: 0xBf4164e07552e4c1b1B1597E72C7acAA3F0a1c58."

      What this check looks for
    Description, as the server wrote it
    Use when you need an exact quote compared with cached public price and recipient evidence. Compare a supplied x402 quote with the latest comparable cached public unpaid quote and disclosed retained history. Scout keeps only the latest probe; the separate public index can retain seven days. $0.001 USDC on Base for a usable comparison, no trial. Missing/stale/ambiguous evidence returns unknown without charge. Never fetches or pays the target. Informational historical comparison, not payment approval or loss prevention. Price: 0.001 USDC. Minimal input example: {"url":"https://api.example.com/data","amount":"10000","payTo":"0x1111111111111111111111111111111111111111","asset":"0x833589fcd6edb6e08f4c7c32d4f71b54bda02913","network":"eip155:8453"}. Illustrative output excerpt: {"status":"unknown","baseline":null,"multiple":null,"suggested_max_amount_per_payment":null,"not_charged":true}. Illustrative output excerpt, not a live result, signature, approval or guarantee. Freshness/data/availability checks still apply. How to pay (MCP): Unknown evidence returns without charge; a usable comparison requires payment. No shared trial. Use an x402-capable MCP client. A generic MCP connector without payment support cannot complete a paid continuation. On payment required, STOP. Inspect the exact accepts amount, asset, network and payTo; obtain explicit wallet/budget approval before continuing. Only after approval, let the configured x402 MCP client attach its signed payload in _meta["x402/payment"] and retry the same tools/call arguments. Never send a separate USDC transfer or paste keys into tool arguments. Never silently fall back from free access to payment. If a signed attempt has an unknown outcome, inspect that authorization/transaction; do not create another automatically.

    Annotations the server declares

    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: false
    • openWorldHint: false
  • security_checkUse when you need the full passive evidence from a current complete saved Lint report.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 when you need the full passive evidence from a current complete saved Lint report. x402 Security Check (Lint Pro). Detailed passive findings and source-linked fixes from an existing, unexpired free Lint report. $0.10 USDC on Base, no trial or automatic payment. No new target request. Resource binding is advertised only; paid-response caching, replay/idempotency, header parser behavior and settlement-before-delivery remain unknown. Not an audit or safety guarantee. Price: 0.10 USDC. Minimal input example: {"report_id":"00000000-0000-4000-8000-000000000000"}. Illustrative output excerpt: {"mode":"passive_unpaid","assessment_complete":false,"unknown_count":5,"methodology":{"target_requests_added":0,"payment_tests":0}}. Illustrative output excerpt, not a live result, signature, approval or guarantee. Freshness/data/availability checks still apply. How to pay (MCP): Inspect the free Lint security summary first; a current complete report_id is required for the paid detail. No automatic refresh or shared trial. Use an x402-capable MCP client. A generic MCP connector without payment support cannot complete a paid continuation. On payment required, STOP. Inspect the exact accepts amount, asset, network and payTo; obtain explicit wallet/budget approval before continuing. Only after approval, let the configured x402 MCP client attach its signed payload in _meta["x402/payment"] and retry the same tools/call arguments. Never send a separate USDC transfer or paste keys into tool arguments. Never silently fall back from free access to payment. If a signed attempt has an unknown outcome, inspect that authorization/transaction; do not create another automatically.

    Annotations the server declares

    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: false
    • openWorldHint: false
  • seller_indexUse when you need sellers from the existing cached public catalog.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 when you need sellers from the existing cached public catalog. Search cached public x402 sellers. First three matches free; up to12 cost0.01 USDC only with extra matches. Bounded unpaid availability, 402, quote history and public Lint evidence; missing evidence unknown. Partial Base USDC wash-activity heuristic: low/elevated/unknown, coverage and reasons. Elevated subtracts10 ordering points without excluding sellers. No target request, safety or paid-delivery guarantee. Price: 0.01 USDC. Minimal input example: {"capability":"weather forecast","limit":3}. Illustrative output excerpt: {"results":[],"free_top":3,"paid_available":false,"not_charged":true}. Illustrative output excerpt, not a live result, signature, approval or guarantee. Freshness/data/availability checks still apply. How to pay (MCP): Top three cached matches are free; extended results cost the quoted fee only when extra matches exist. No shared trial. Use an x402-capable MCP client. A generic MCP connector without payment support cannot complete a paid continuation. On payment required, STOP. Inspect the exact accepts amount, asset, network and payTo; obtain explicit wallet/budget approval before continuing. Only after approval, let the configured x402 MCP client attach its signed payload in _meta["x402/payment"] and retry the same tools/call arguments. Never send a separate USDC transfer or paste keys into tool arguments. Never silently fall back from free access to payment. If a signed attempt has an unknown outcome, inspect that authorization/transaction; do not create another automatically.

    Annotations the server declares

    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: false
    • openWorldHint: false
  • size_solarUse when you need solar energy arithmetic from supplied assumptions; see Data APIs.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 when you need solar energy arithmetic from supplied assumptions; see Data APIs. Off-grid PV energy-balance estimate based on Wattpath by Nomad Energy Club. Supply daily loads, peak sun hours, PV losses, and optional battery assumptions. No location/weather lookup, electrical design, surge or installation certification. $0.02 per calculation; no stored inputs. Price: 0.02 USDC. Minimal input example: {"loads":[{"watts":100,"hours_per_day":5,"quantity":1,"kind":"dc"}],"peak_sun_hours":4,"pv_loss_fraction":0.2,"panel_watts":200,"battery":{"voltage":12,"usable_fraction":0.8,"autonomy_days":2}}. Illustrative output excerpt: {"daily_energy_wh":{"battery_side":400},"pv":{"minimum_nameplate_watts":125,"rounded_up_watts":125},"assumptions":{"energy_balance_only":true,"weather_or_location_lookup":false}}. Illustrative output excerpt, not a live result, signature, approval or guarantee. Freshness/data/availability checks still apply. How to pay (MCP): First use the existing shared allowance of 3 eligible calls per IP per UTC day, without payment metadata. This is shared with eligible HTTP trials, not a new allowance. Use an x402-capable MCP client. A generic MCP connector without payment support cannot complete a paid continuation. On payment required, STOP. Inspect the exact accepts amount, asset, network and payTo; obtain explicit wallet/budget approval before continuing. Only after approval, let the configured x402 MCP client attach its signed payload in _meta["x402/payment"] and retry the same tools/call arguments. Never send a separate USDC transfer or paste keys into tool arguments. Never silently fall back from free access to payment. If a signed attempt has an unknown outcome, inspect that authorization/transaction; do not create another automatically.

    Annotations the server declares

    • readOnlyHint: true
    • openWorldHint: false
  • verify_payee_bindingUse when a seller advertises an ERC-8004 identity and you need an ERC-8478 Draft wallet-binding observation.Reads text others write3 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.

    Findings

    • Warn

      The description contains a 42-character base64 string.

      Matched text

      description: "...p155:8453","asset":"0x833589fCD6eDb6E08f4c7C32D4f71b54bdA02913","amount":"10000","..."

      What this check looks for
    • Warn

      The description contains a 42-character base64 string.

      Matched text

      description: "...t":"10000","payTo":"0xBf4164e07552e4c1b1B1597E72C7acAA3F0a1c58"},"resource_url":"h..."

      What this check looks for
    • Warn

      The input schema at properties.requirement.properties.payTo.description contains a 42-character base64 string.

      Matched text

      inputSchema.properties.requirement.properties.payTo.description: "...e address. Example: 0xBf4164e07552e4c1b1B1597E72C7acAA3F0a1c58."

      What this check looks for
    Description, as the server wrote it
    Use when a seller advertises an ERC-8004 identity and you need an ERC-8478 Draft wallet-binding observation. ERC-8478 Draft payee verification: endpoint origin and ERC-8004 agent wallet or EIP-712 delegation at one pinned registry block. Nine explicit outcomes; not-read is uncharged. Signed JSON observation, not payment approval or a safety guarantee. Shared trial may apply. Price: 0.01 USDC. Minimal input example: {"requirement":{"scheme":"exact","network":"eip155:8453","asset":"0x833589fCD6eDb6E08f4c7C32D4f71b54bdA02913","amount":"10000","payTo":"0xBf4164e07552e4c1b1B1597E72C7acAA3F0a1c58"},"resource_url":"https://api.spendpreflight.com/v1/check"}. Illustrative output excerpt: {"result":{"kind":"payee_binding","outcome":"no-identity","block_number":null,"block_timestamp":null},"jws":"illustrative-not-a-live-signature"}. Illustrative output excerpt, not a live result, signature, approval or guarantee. Freshness/data/availability checks still apply. How to pay (MCP): First use the existing shared allowance of 3 eligible calls per IP per UTC day, without payment metadata. This is shared with eligible HTTP trials, not a new allowance. Use an x402-capable MCP client. A generic MCP connector without payment support cannot complete a paid continuation. On payment required, STOP. Inspect the exact accepts amount, asset, network and payTo; obtain explicit wallet/budget approval before continuing. Only after approval, let the configured x402 MCP client attach its signed payload in _meta["x402/payment"] and retry the same tools/call arguments. Never send a separate USDC transfer or paste keys into tool arguments. Never silently fall back from free access to payment. If a signed attempt has an unknown outcome, inspect that authorization/transaction; do not create another automatically.

    Annotations the server declares

    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: false
    • openWorldHint: true
  • wallet_statementUse when you need a reconciled Base USDC statement for a bounded UTC window.Reads text others write3 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.

    Findings

    • Warn

      The description contains a 42-character base64 string.

      Matched text

      description: "...xample: {"address":"0xBf4164e07552e4c1b1B1597E72C7acAA3F0a1c58","start":"2026-10-2..."

      What this check looks for
    • Warn

      The input schema at properties.address.description contains a 42-character base64 string.

      Matched text

      inputSchema.properties.address.description: "...ss string. Example: 0xBf4164e07552e4c1b1B1597E72C7acAA3F0a1c58."

      What this check looks for
    • Warn

      The input schema at description contains a 42-character base64 string.

      Matched text

      inputSchema.description: "...uments: {"address":"0xBf4164e07552e4c1b1B1597E72C7acAA3F0a1c58","start":"2026-10-2..."

      What this check looks for
    Description, as the server wrote it
    Use when you need a reconciled Base USDC statement for a bounded UTC window. Export native Base USDC wallet transfers as JSON or CSV for a past range up to 90 days, start inclusive/end exclusive. Exact decimal running balances, counterparties and OFAC flags. Up to 1000 transfers/128 RPC calls/25 seconds; oversized or unavailable historical data fails without charge. Informational, not tax advice. Price $0.05 USDC on Base. No free trial. Price: 0.05 USDC. Minimal input example: {"address":"0xBf4164e07552e4c1b1B1597E72C7acAA3F0a1c58","start":"2026-10-22T22:07:00Z","end":"2026-10-22T22:08:00Z","format":"json"}. Illustrative output excerpt: {"kind":"wallet_statement","opening_balance":"4.188374","closing_balance":"4.238374","complete":true,"rows":[{"direction":"in","amount":"0.050000","running_balance":"4.238374"}]}. Illustrative output excerpt, not a live result, signature, approval or guarantee. Freshness/data/availability checks still apply. How to pay (MCP): No shared free trial for this tool. Use an x402-capable MCP client. A generic MCP connector without payment support cannot complete a paid continuation. On payment required, STOP. Inspect the exact accepts amount, asset, network and payTo; obtain explicit wallet/budget approval before continuing. Only after approval, let the configured x402 MCP client attach its signed payload in _meta["x402/payment"] and retry the same tools/call arguments. Never send a separate USDC transfer or paste keys into tool arguments. Never silently fall back from free access to payment. If a signed attempt has an unknown outcome, inspect that authorization/transaction; do not create another automatically.

    Annotations the server declares

    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: false
    • openWorldHint: true
Only reads 2Show 2 read-only tools
  • find_rv_stationsUse when you need nearby RV stations from the dated FillPoint snapshot; see Data APIs.No findings

    Why it is in this group

    • The name "find_rv_stations" says "find": a word for reading.
    Description, as the server wrote it
    Use when you need nearby RV stations from the dated FillPoint snapshot; see Data APIs. Nearest RV dump/water stations from FillPoint RV's dated September 17, 2026 ODbL snapshot. $0.01 for bounded straight-line proximity calculation, not live access, potable-water assurance or road routing. Snapshot age and unknown current conditions are explicit. Full underlying projection is free under ODbL. Price: 0.01 USDC. Minimal input example: {"latitude":42.07,"longitude":-88.36,"radius_km":50,"limit":10,"service":"dump"}. Illustrative output excerpt: {"results":[],"distance_method":"great_circle_not_road_distance"}. Illustrative output excerpt, not a live result, signature, approval or guarantee. Freshness/data/availability checks still apply. How to pay (MCP): First use the existing shared allowance of 3 eligible calls per IP per UTC day, without payment metadata. This is shared with eligible HTTP trials, not a new allowance. Use an x402-capable MCP client. A generic MCP connector without payment support cannot complete a paid continuation. On payment required, STOP. Inspect the exact accepts amount, asset, network and payTo; obtain explicit wallet/budget approval before continuing. Only after approval, let the configured x402 MCP client attach its signed payload in _meta["x402/payment"] and retry the same tools/call arguments. Never send a separate USDC transfer or paste keys into tool arguments. Never silently fall back from free access to payment. If a signed attempt has an unknown outcome, inspect that authorization/transaction; do not create another automatically.

    Annotations the server declares

    • readOnlyHint: true
    • openWorldHint: false
  • get_service_infoUse when you need more tools: browse Buyers, Sellers, Books and Data APIs, prices, rules and examples.Reads text others writeNo findings

    Why it is in this group

    • The name "get_service_info" says "get": a word for reading.

    What it can reach

    • untrusted-inputThe description's first sentence says "browse": it reads web pages, which anyone can write.
    Description, as the server wrote it
    Use when you need more tools: browse Buyers, Sellers, Books and Data APIs, prices, rules and examples. If you have a payment challenge, call preflight_payment directly. Free, read-only service reference. Use when you need prices, default spending rules, payment setup, trial allowance, or a ready-to-use example call. If you already have a payment challenge or cart, call preflight_payment directly; this reference call is optional. Takes no arguments and performs no payee screening or payment decision. Use check_payee for standalone payee risk; use preflight_payment for an x402 challenge or cart and an allow/hold/block decision.

    Annotations the server declares

    • readOnlyHint: true
    • 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.

No private-data, untrusted-input, outbound path in this server alone: no tool is tagged private-data.

Reads private data0
Reads files, mail, messages, databases, secrets, calendars or repositories.
Reads text others write11
Brings in text someone else can write: web pages, inbound email, messages, issues and comments.
Can send data out5
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.

Tools listed without sign-in

Looked at Nov 20, 2026, 19:11 UTC, unchanged since Nov 6, 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
15
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 15 tools that change things (2 irreversible, 13 write). Whether a call also works without credentials was not tested: tools are never called.

    Matched text

    https://api.spendpreflight.com/mcp: tools/list returned 17 tools to an anonymous client, including preflight_payment (irreversible), payment_receipt (irreversible), check_payee (write), lint_endpoint (write), check_token (write) and 10 more

    What this check looks for

Posture changes

  1. Nov 6, 2026
    • Tools that change things, listed without credentials: 12 to 15
  2. Oct 29, 2026
    • Tools that change things, listed without credentials: 6 to 12

Findings

Matches of fixed checks, with the text they matched.

17 warn11 info
  • The input schema at properties.challenge.description contains a 42-character base64 string.

    Matched text

    inputSchema.properties.challenge.description: "...4bda02913","payTo":"0xBf4164e07552e4c1b1B1597E72C7acAA3F0a1c58","amount":"10000","..."

    What this check looks for
  • The input schema at properties.cart.properties.pay_to.description contains a 42-character base64 string.

    Matched text

    inputSchema.properties.cart.properties.pay_to.description: "...ss string. Example: 0xBf4164e07552e4c1b1B1597E72C7acAA3F0a1c58."

    What this check looks for
  • The input schema at properties.cart.description contains a 42-character base64 string.

    Matched text

    inputSchema.properties.cart.description: "...usd":1.25,"pay_to":"0xBf4164e07552e4c1b1B1597E72C7acAA3F0a1c58"}. total_usd is a U..."

    What this check looks for
  • The server marks "preflight_payment" read-only (readOnlyHint: true), and its own name, description or input give the class "irreversible".

    Matched text

    annotations: {"readOnlyHint":true,"idempotentHint":false,"openWorldHint":true}; The name "preflight_payment" says "payment": a word that usually means money moves, which cannot be taken back.

    What this check looks for
  • The input schema at properties.address.description contains a 42-character base64 string.

    Matched text

    inputSchema.properties.address.description: "...Base-only. Example: 0xBf4164e07552e4c1b1B1597E72C7acAA3F0a1c58. At least one of do..."

    What this check looks for
  • The description contains a 42-character base64 string.

    Matched text

    description: "...xample: {"address":"0x833589fCD6eDb6E08f4c7C32D4f71b54bdA02913","chain":"eip155:84..."

    What this check looks for
Show 22 more findings
  • The input schema at properties.address.description contains a 42-character base64 string.

    Matched text

    inputSchema.properties.address.description: "...ss string. Example: 0xBf4164e07552e4c1b1B1597E72C7acAA3F0a1c58."

    What this check looks for
  • The input schema at description contains a 42-character base64 string.

    Matched text

    inputSchema.description: "...uments: {"address":"0x833589fCD6eDb6E08f4c7C32D4f71b54bdA02913","chain":"eip155:84..."

    What this check looks for
  • The input schema at properties.pay_to.description contains a 42-character base64 string.

    Matched text

    inputSchema.properties.pay_to.description: "...ss string. Example: 0xBf4164e07552e4c1b1B1597E72C7acAA3F0a1c58. Required with amou..."

    What this check looks for
  • The server marks "payment_receipt" read-only (readOnlyHint: true), and its own name, description or input give the class "irreversible".

    Matched text

    annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":false,"openWorldHint":true}; The name "payment_receipt" says "payment": a word that usually means money moves, which cannot be taken back.

    What this check looks for
  • The description contains a 42-character base64 string.

    Matched text

    description: "...p155:8453","asset":"0x833589fCD6eDb6E08f4c7C32D4f71b54bdA02913","amount":"10000","..."

    What this check looks for
  • The description contains a 42-character base64 string.

    Matched text

    description: "...t":"10000","payTo":"0xBf4164e07552e4c1b1B1597E72C7acAA3F0a1c58"},"resource_url":"h..."

    What this check looks for
  • The input schema at properties.requirement.properties.payTo.description contains a 42-character base64 string.

    Matched text

    inputSchema.properties.requirement.properties.payTo.description: "...e address. Example: 0xBf4164e07552e4c1b1B1597E72C7acAA3F0a1c58."

    What this check looks for
  • The input schema at properties.payTo.description contains a 42-character base64 string.

    Matched text

    inputSchema.properties.payTo.description: "...e address. Example: 0xBf4164e07552e4c1b1B1597E72C7acAA3F0a1c58."

    What this check looks for
  • The description contains a 42-character base64 string.

    Matched text

    description: "...xample: {"address":"0xBf4164e07552e4c1b1B1597E72C7acAA3F0a1c58","start":"2026-10-2..."

    What this check looks for
  • The input schema at properties.address.description contains a 42-character base64 string.

    Matched text

    inputSchema.properties.address.description: "...ss string. Example: 0xBf4164e07552e4c1b1B1597E72C7acAA3F0a1c58."

    What this check looks for
  • The input schema at description contains a 42-character base64 string.

    Matched text

    inputSchema.description: "...uments: {"address":"0xBf4164e07552e4c1b1B1597E72C7acAA3F0a1c58","start":"2026-10-2..."

    What this check looks for
  • The description is 3,123 characters long, over the 2,000 this check looks for.

    Matched text

    description ends: "...ome, inspect that authorization/transaction; do not create another automatically."

    What this check looks for
  • The description mentions "credentials".

    Matched text

    description: "...ar acceptance guarantees. No payment or credentials; GET cannot validate a POST-only endpoi..."

    What this check looks for
  • The input schema at properties.url.description mentions "credentials".

    Matched text

    inputSchema.properties.url.description: "...resource URL string, including path; no credentials. Example: https://api.example.com/data."

    What this check looks for
  • The description is 2,118 characters long, over the 2,000 this check looks for.

    Matched text

    description ends: "...ome, inspect that authorization/transaction; do not create another automatically."

    What this check looks for
  • The input schema at properties.url.description mentions "credentials".

    Matched text

    inputSchema.properties.url.description: "...resource URL string, including path; no credentials. Use the current published example in O..."

    What this check looks for
  • The input schema at properties.url.description mentions "credentials".

    Matched text

    inputSchema.properties.url.description: "...resource URL string, including path; no credentials. Example: https://api.example.com/data."

    What this check looks for
  • The input schema at properties.url.description mentions "credentials".

    Matched text

    inputSchema.properties.url.description: "...resource URL string, including path; no credentials. Example: https://api.example.com/data."

    What this check looks for
  • The description is 2,076 characters long, over the 2,000 this check looks for.

    Matched text

    description ends: "...ome, inspect that authorization/transaction; do not create another automatically."

    What this check looks for
  • The input schema at properties.domain.description mentions "credentials".

    Matched text

    inputSchema.properties.domain.description: "...ing without scheme, path, IP address or credentials. Example: example.com."

    What this check looks for
  • The input schema at properties.url.description mentions "credentials".

    Matched text

    inputSchema.properties.url.description: "...resource URL string, including path; no credentials. Example: https://api.example.com/data."

    What this check looks for
  • The description is 2,021 characters long, over the 2,000 this check looks for.

    Matched text

    description ends: "...ome, inspect that authorization/transaction; do not create another automatically."

    What this check looks for

History

Changes to its tool definitions and to its remote endpoint, newest first.

  1. Tools3 changed.~3 changedView diff
  2. PostureRemote posture changed. Tools that change things, listed without credentials: 12 -> 15.View diff
    • Tools that change things, listed without credentials: 12 to 15
  3. Tools16 changed.~16 changedView diff
  4. Tools17 changed.~17 changedView diff
  5. Tools1 added, 14 changed.+1 added~14 changedView diff
  6. PostureRemote posture changed. Tools that change things, listed without credentials: 6 -> 12.View diff
    • Tools that change things, listed without credentials: 6 to 12
  7. Tools4 added, 1 changed.+4 added~1 changedView diff
  8. Tools2 added.+2 addedView diff
  9. Tools1 changed.~1 changedView diff
  10. Tools5 added, 1 changed.+5 added~1 changedView diff

Install

Pinned to version 1.2.0-metadata.9. Values in angle brackets are yours to fill in.

claude mcp add --transport http spendpreflight https://api.spendpreflight.com/mcp
claude mcp add --transport http spendpreflight https://api.spendpreflight.com/mcp

Run 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.

.claude/settings.json
{
  "permissions": {
    "allow": [
      "mcp__spendpreflight__get_service_info",
      "mcp__spendpreflight__find_rv_stations"
    ],
    "ask": [
      "mcp__spendpreflight__preflight_payment",
      "mcp__spendpreflight__check_payee",
      "mcp__spendpreflight__lint_endpoint",
      "mcp__spendpreflight__check_token",
      "mcp__spendpreflight__payment_receipt",
      "mcp__spendpreflight__verify_payee_binding",
      "mcp__spendpreflight__quote_check",
      "mcp__spendpreflight__listing_guard",
      "mcp__spendpreflight__check_listing",
      "mcp__spendpreflight__security_check",
      "mcp__spendpreflight__seller_index",
      "mcp__spendpreflight__domain_auth",
      "mcp__spendpreflight__check_endpoint",
      "mcp__spendpreflight__wallet_statement",
      "mcp__spendpreflight__size_solar"
    ],
    "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 spendpreflight. Review it before you use it.

Capability tags behind 13 rules
  • mcp__spendpreflight__preflight_payment untrusted-input
  • mcp__spendpreflight__check_payee untrusted-input
  • mcp__spendpreflight__lint_endpoint untrusted-input, outbound
  • mcp__spendpreflight__check_token untrusted-input
  • mcp__spendpreflight__payment_receipt untrusted-input
  • mcp__spendpreflight__get_service_info untrusted-input
  • mcp__spendpreflight__verify_payee_binding untrusted-input
  • mcp__spendpreflight__quote_check outbound
  • mcp__spendpreflight__listing_guard untrusted-input, outbound
  • mcp__spendpreflight__check_listing untrusted-input, outbound
  • mcp__spendpreflight__domain_auth untrusted-input
  • mcp__spendpreflight__check_endpoint outbound
  • mcp__spendpreflight__wallet_statement 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.spendpreflight/spendpreflight/verify?hash=fbbaad0b56c427ea0a155791ef5de13127da79815e20a40c8981c5be69f4270b" | jq -e '.changed == false'

Pinned hash fbbaad0b56c4, from Nov 15, 2026, 20:18 UTC.

Watch

Get told when its tools change: a new tool, a new description, a new class.

https://radar.anlyon.com/servers/io.github.spendpreflight/spendpreflight/changes.xml

An Atom feed. Works in any feed reader or Slack's RSS app.

Badge

Anlyon Radar badge for io.github.spendpreflight/spendpreflight
[![Anlyon Radar](https://radar.anlyon.com/badge/io.github.spendpreflight/spendpreflight.svg)](https://radar.anlyon.com/servers/io.github.spendpreflight/spendpreflight)

Source: MCP Registry · repository. Synced Nov 15, 2026, 20:18 UTC.

SpendPreflight MCP server: questions and answers

Answered from the scan above, not from the publisher's description.

Is the SpendPreflight MCP server safe?

No scan can promise that. What its definitions show: 17 tools, 15 of which can change, send or delete something, and 28 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 SpendPreflight MCP server delete data?

No tool is classed as destructive: going by their definitions, none of them deletes, cancels or revokes anything.

Can the SpendPreflight MCP server do things that cannot be undone?

Yes. 2 tools can send, pay, publish or merge: preflight_payment and payment_receipt.

What tools does the SpendPreflight MCP server have?

17 tools: 2 irreversible, 13 write, 2 read. preflight_payment, payment_receipt, check_payee, lint_endpoint, check_token, verify_payee_binding, quote_check, listing_guard and 9 more.

Does the SpendPreflight MCP server need sign-in or OAuth?

Its remote endpoint listed its tools to an anonymous client, without sign-in.

Does the SpendPreflight MCP server send data to outside services?

5 tools can carry data out (it sends, posts or publishes, or takes a recipient, channel or URL): lint_endpoint, quote_check, listing_guard, check_listing and check_endpoint.

Has the SpendPreflight MCP server changed its tools?

Yes, 8 times since Anlyon Radar started tracking it, most recently on Nov 15, 2026. Every change is on this page and in its Atom feed.