Kit Rx MCP server
com.kit-rx.api/kitrxapi.kit-rx.comstreamable-httpv1.3.0Scanned
Pay-per-call AI briefs and translation, web markdown, WHOIS/DNS, SEC, onchain safety (x402 USDC)
49 tools: 3 do something that can't be undone, 40 change things and 6 only read.
- Reads private data · 2
- Reads text others write · 48
- Can send data out · 12
- 26 findings
Tools listed without sign-in. Remote posture
With Anlyon
3 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.
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 Kit Rx MCP server can do
The Kit Rx MCP server is published by api.kit-rx.com, a domain the MCP Registry verified. It has 49 tools. It can do things that cannot be undone, such as sending, paying or publishing (x402_preflight, chain_calldata_decode and text_translate) and create or change data (40 tools: rx_drug, rx_shortages, rx_recalls, rx_ndc_convert and 36 more), but cannot delete data. 6 tools only read.
2 tools read private data (x402_find and web_read); 48 tools bring in text other people can write (rx_drug, rx_shortages, rx_recalls and 45 more); 12 tools can send data out of your system (web_check, web_headers, web_markdown and 9 more). 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 26 findings (annotation-conflict, credentials-mention and base64-blob). Each one comes with the text it matched.
Its remote endpoint lists its tools without sign-in, and its tools changed 2 times since Anlyon Radar first saw them.
Last scanned Nov 16, 2026, version 1.3.0.
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.
49 tools
Can't be undone 3Claude Code: ask
chain_calldata_decodeDecodes EVM calldata before signing: function signature, selector, named arguments, candidate signatures for ambiguous selectors, one level of bundled calls (multicall, Multicall3,…Reads text others writeCan send data out2 findings
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
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
- outboundThe input takes a destination (input.to): what the tool writes goes there.
Findings
- Warn
The input schema at properties.to.examples[0] contains a 42-character base64 string.
What this check looks forMatched text
inputSchema.properties.to.examples[0]: "0x833589fCD6eDb6E08f4c7C32D4f71b54bdA02913"
- Warn
The server marks "chain_calldata_decode" read-only (readOnlyHint: true), and its own name, description or input give the class "irreversible".
What this check looks forMatched text
annotations: {"readOnlyHint":true,"openWorldHint":true}; The input addresses people (input.to): what it sends cannot be taken back.
Description, as the server wrote it Decodes EVM calldata before signing: function signature, selector, named arguments, candidate signatures for ambiguous selectors, one level of bundled calls (multicall, Multicall3, ERC-4337 execute, Safe execTransaction, multiSend), token symbol and decimals, and risk flags with reasons for unlimited approvals, setApprovalForAll, permits, transferFrom, admin calls and delegatecall. Sources: Sourcify ABIs and 4byte signatures, JSON-RPC via PublicNode. Price: $0.002 per successful call, paid in USDC with x402 v2; failed calls and calls that match nothing are not charged.
Annotations the server declares
- title: EVM calldata decoder with pre-signing risk flags
- readOnlyHint: true
- openWorldHint: true
text_translateAI translation of up to 8,000 characters into a target language given as an ISO 639-1 code or English name (with optional region or script, e.g.Reads text others writeCan 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
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
- outboundThe input takes a destination (input.to): what the tool writes goes there.
Findings
- Warn
The server marks "text_translate" read-only (readOnlyHint: true), and its own name, description or input give the class "irreversible".
What this check looks forMatched text
annotations: {"readOnlyHint":true,"openWorldHint":true}; The input addresses people (input.to): what it sends cannot be taken back.
Description, as the server wrote it AI translation of up to 8,000 characters into a target language given as an ISO 639-1 code or English name (with optional region or script, e.g. pt-BR, zh-Hant). Returns the translated text, detected source language, target language, character counts and a truncation flag. Machine translation by Anthropic Claude under a fixed Kit Rx prompt that asks it to keep formatting, numbers, URLs and names; may contain errors; not a certified translation. Price: $0.005 per successful call, paid in USDC with x402 v2; failed calls and calls that match nothing are not charged. AI-generated answer; a call with no arguments returns the stored example answer without a model call.
Annotations the server declares
- title: AI text translation into 180+ languages
- readOnlyHint: true
- openWorldHint: true
x402_preflightSends an unpaid GET to an x402 endpoint and reports what it would charge before paying: HTTP status, redirects (max 3, re-validated), v2 PAYMENT-REQUIRED header or v1 body challeng…Reads text others writeCan send data out2 findings
Why it is in this group
- The description's first verb is "send": a word for sending, and what is sent 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.
- outboundThe description's first verb is "send": it sends or publishes something to other people.
- 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".
What this check looks forMatched text
inputSchema.properties.url.description: "...racters. Default port only; no embedded credentials, IP-address hosts, localhost or private..."
- Warn
The server marks "x402_preflight" read-only (readOnlyHint: true), and its own name, description or input give the class "irreversible".
What this check looks forMatched text
annotations: {"readOnlyHint":true,"openWorldHint":true}; The description's first verb is "send": a word for sending, and what is sent cannot be taken back.
Description, as the server wrote it Sends an unpaid GET to an x402 endpoint and reports what it would charge before paying: HTTP status, redirects (max 3, re-validated), v2 PAYMENT-REQUIRED header or v1 body challenge and its validity, each accepts entry (scheme, CAIP-2 network/name, atomic amount, USD for canonical USDC, asset check, payTo, timeout, EIP-712 name/version), OFAC SDN payTo screen, resource, Bazaar extension, risk flags, host context from registry RDAP (registrar, domain age) and Cloudflare DoH. Price: $0.005 per successful call, paid in USDC with x402 v2; failed calls and calls that match nothing are not charged.
Annotations the server declares
- title: x402 preflight: inspect a paid endpoint before paying
- readOnlyHint: true
- openWorldHint: true
Changes things 40Claude Code: ask
chain_gasLive gas fees and estimated USD transaction costs on Base, Ethereum, Arbitrum, OP Mainnet, Polygon and Avalanche: base fee and next-block estimate, priority fee p25/p50/p75 over 10…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 Live gas fees and estimated USD transaction costs on Base, Ethereum, Arbitrum, OP Mainnet, Polygon and Avalanche: base fee and next-block estimate, priority fee p25/p50/p75 over 10 blocks, gas price, 10-block base-fee trend, suggested max fee, and USD cost of a native transfer, ERC-20 transfer and swap (upper-bound rollup L1 fees included), plus the cheapest chain per action. Sources: JSON-RPC via PublicNode (public RPC fallback); prices from on-chain Uniswap v3 pools. Price: $0.002 per successful call, paid in USDC with x402 v2; failed calls and calls that match nothing are not charged.
Annotations the server declares
- title: EVM gas fees and USD transaction costs (6 chains)
- readOnlyHint: true
- openWorldHint: true
company_briefCompany due-diligence brief for a domain, name or ticker: Wikidata and GLEIF identity (legal name, LEI, HQ country, website, founding), SEC EDGAR ticker, CIK, periodic filings, 8-K…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 Company due-diligence brief for a domain, name or ticker: Wikidata and GLEIF identity (legal name, LEI, HQ country, website, founding), SEC EDGAR ticker, CIK, periodic filings, 8-K events (last 180 days), XBRL revenue, net income, assets and cash copied from source, OFAC sanctions screening, domain registration and risk flags. Overview and risk notes are AI-generated by Anthropic Claude from that data and the homepage text; not investment or legal advice. Sources listed. Price: $0.04 per successful call, paid in USDC with x402 v2; failed calls and calls that match nothing are not charged. AI-generated answer; a call with no arguments returns the stored example answer without a model call.
Annotations the server declares
- title: AI company brief: identity, SEC, sanctions, domain risk
- readOnlyHint: true
- openWorldHint: true
company_brief_proAI company research brief from SEC filings: the latest 10-K (business, risk factors, MD&A), recent 8-Ks and XBRL financials.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 AI company research brief from SEC filings: the latest 10-K (business, risk factors, MD&A), recent 8-Ks and XBRL financials. Returns business model, segments, competitive position, key risks each with a verbatim 10-K quote checked in code and an EDGAR link, recent developments, watch items, and a financial snapshot copied from XBRL. Lookup by ticker, name, domain or LEI. AI-generated by Anthropic Claude; companies without a 10-K get a profile and website-only brief. Price: $0.25 per successful call, paid in USDC with x402 v2; failed calls and calls that match nothing are not charged. AI-generated answer; a call with no arguments returns the stored example answer without a model call.
Annotations the server declares
- title: AI company brief from 10-K risk factors, 8-Ks and XBRL
- readOnlyHint: true
- openWorldHint: true
company_profileCompany card for a domain or name.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 Company card for a domain or name. From Wikidata: QID, label, description, websites, inception, industries, headquarters, country, latest employee count, parent organizations, stock listings and tickers, founders (names only), legal form, LEI. From GLEIF when an LEI is known or a legal name matches exactly: legal name, jurisdiction, entity and registration status, legal and HQ city/country, parents. With match method and confidence. Sources: Wikidata, GLEIF (CC0). Price: $0.01 per successful call, paid in USDC with x402 v2; failed calls and calls that match nothing are not charged.
Annotations the server declares
- title: Company profile from Wikidata and LEI records
- readOnlyHint: true
- openWorldHint: true
company_reportAI KYB and vendor due-diligence report on a company from its domain, name or US ticker: identity, GLEIF ownership (parents, subsidiaries), business, SEC XBRL financials copied from…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 AI KYB and vendor due-diligence report on a company from its domain, name or US ticker: identity, GLEIF ownership (parents, subsidiaries), business, SEC XBRL financials copied from filings, recent 8-K events, domain registration, email DNS, website security headers and OFAC screening of the entity and related companies, plus risk ratings, red flags and open questions, quotes checked against cited sources. Written by Anthropic Claude; takes 1-2 minutes; not investment advice. Price: $0.75 per successful call, paid in USDC with x402 v2; failed calls and calls that match nothing are not charged. AI-generated answer; a call with no arguments returns the stored example answer without a model call.
Annotations the server declares
- title: AI company due-diligence report: KYB, ownership, risk
- readOnlyHint: true
- openWorldHint: true
compliance_sanctionsOFAC sanctions (AML/KYC) screening of a crypto wallet or a person, company, vessel or aircraft name against a dated snapshot of the U.S.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.examples[0] contains a 42-character base64 string.
What this check looks forMatched text
inputSchema.properties.address.examples[0]: "0x098B716B8Aaf21512996dC57EB0615e2383E2f96"
Description, as the server wrote it OFAC sanctions (AML/KYC) screening of a crypto wallet or a person, company, vessel or aircraft name against a dated snapshot of the U.S. Treasury SDN and non-SDN Consolidated lists. Address mode: exact match on OFAC-listed digital currency addresses (BTC, ETH/EVM, USDT, TRX, XMR, LTC, more). Name mode: fuzzy match on names and strong aliases ignoring case, accents, punctuation and word order, scored 0-100. Returns match flag, OFAC uid, type, programs, country, list date. Price: $0.01 per successful call, paid in USDC with x402 v2; failed calls and calls that match nothing are not charged.
Annotations the server declares
- title: OFAC sanctions screening: crypto wallet or name
- readOnlyHint: true
- openWorldHint: true
contract_abiABI and metadata for an EVM contract on Base, Ethereum, Arbitrum, OP Mainnet, Polygon or Avalanche: Sourcify verification match, contract name, compiler, proxy type and implementat…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.examples[0] contains a 42-character base64 string.
What this check looks forMatched text
inputSchema.properties.address.examples[0]: "0x833589fCD6eDb6E08f4c7C32D4f71b54bdA02913"
Description, as the server wrote it ABI and metadata for an EVM contract on Base, Ethereum, Arbitrum, OP Mainnet, Polygon or Avalanche: Sourcify verification match, contract name, compiler, proxy type and implementations (implementation ABI merged), function, event and error signatures with selectors and topics, and read vs write function lists. Unverified contracts get a best-effort function list from bytecode selectors named via the Sourcify 4byte database. Sources: Sourcify, chain JSON-RPC via PublicNode. Price: $0.002 per successful call, paid in USDC with x402 v2; failed calls and calls that match nothing are not charged.
Annotations the server declares
- title: EVM contract ABI, proxy info and signatures (6 chains)
- readOnlyHint: true
- openWorldHint: true
doc_markdownMarkdown of a document at a public URL, up to 10 MB: PDF (page count, title, created and modified dates from its metadata), DOCX, ODT, XLSX/XLS, ODS, Numbers (sheet names), CSV, XM…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".
What this check looks forMatched text
inputSchema.properties.url.description: "...le (at most 10 MB). Default ports only; credentials, IP addresses, localhost and private or..."
Description, as the server wrote it Markdown of a document at a public URL, up to 10 MB: PDF (page count, title, created and modified dates from its metadata), DOCX, ODT, XLSX/XLS, ODS, Numbers (sheet names), CSV, XML or HTML; plain text is returned as served. Returns file name, MIME type, size, Markdown, word count, token estimate and truncation flag. Images are not converted. Conversion: Cloudflare Workers AI toMarkdown; fetch via validated redirects, DNS checks via Cloudflare DoH. Price: $0.005 per successful call, paid in USDC with x402 v2; failed calls and calls that match nothing are not charged.
Annotations the server declares
- title: Document (PDF, DOCX, XLSX, CSV) to Markdown
- readOnlyHint: true
- openWorldHint: true
doc_ocrAI OCR of a scanned PDF (up to 10 pages, 5 MB) or a PNG, JPEG or WebP image (up to 7.5 MB) at a public URL: printed text, tables, headings, lists, checkboxes and legible handwritin…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".
What this check looks forMatched text
inputSchema.properties.url.description: "...nd 8000 x 8000 px). Default ports only; credentials, IP addresses, localhost and private or..."
Description, as the server wrote it AI OCR of a scanned PDF (up to 10 pages, 5 MB) or a PNG, JPEG or WebP image (up to 7.5 MB) at a public URL: printed text, tables, headings, lists, checkboxes and legible handwriting transcribed to Markdown page by page, with page count, file type, size and image dimensions. For scans, photos, screenshots, receipts and forms without a text layer. AI-generated by Anthropic Claude under a fixed Kit Rx prompt; transcription can misread characters. URL fetch is SSRF-guarded. Price: $0.02 per successful call, paid in USDC with x402 v2; failed calls and calls that match nothing are not charged. AI-generated answer; a call with no arguments returns the stored example answer without a model call.
Annotations the server declares
- title: AI OCR: scanned PDF or image to Markdown
- readOnlyHint: true
- openWorldHint: true
domain_dnsCurrent DNS records for a hostname from Cloudflare DNS over HTTPS: A, AAAA, CNAME chain, MX with priority and null-MX flag, NS, TXT (up to 25 records of at most 400 characters, ema…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".
What this check looks forMatched text
inputSchema.properties.domain.description: "...address or URL. Scheme, path, port and credentials are stripped and IDNs are converted to..."
Description, as the server wrote it Current DNS records for a hostname from Cloudflare DNS over HTTPS: A, AAAA, CNAME chain, MX with priority and null-MX flag, NS, TXT (up to 25 records of at most 400 characters, email addresses redacted), CAA as flag, tag and value, and SOA primary nameserver, serial and timers, each with its minimum TTL. An optional type parameter limits the answer to one record type. For subdomains without their own zone, NS and SOA come from the registrable domain. Price: $0.003 per successful call, paid in USDC with x402 v2; failed calls and calls that match nothing are not charged.
Annotations the server declares
- title: DNS records for a domain
- readOnlyHint: true
- openWorldHint: true
domain_email_authEmail authentication posture of a domain from DNS (Cloudflare DNS over HTTPS): MX records and null MX, SPF parsed into mechanisms with its all qualifier and a recursive DNS-lookup …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".
What this check looks forMatched text
inputSchema.properties.domain.description: "...address or URL. Scheme, path, port and credentials are stripped and IDNs are converted to..."
Description, as the server wrote it Email authentication posture of a domain from DNS (Cloudflare DNS over HTTPS): MX records and null MX, SPF parsed into mechanisms with its all qualifier and a recursive DNS-lookup count against the RFC 7208 limit of 10, DMARC policy (p, sp, pct, alignment, report addresses' domains), MTA-STS and TLS-RPT records, and BIMI. DKIM is not assessed because selectors cannot be enumerated. Returns an A-F grade and 0-100 score with weighted reasons. Price: $0.003 per successful call, paid in USDC with x402 v2; failed calls and calls that match nothing are not charged.
Annotations the server declares
- title: Email authentication posture (SPF, DMARC, MTA-STS)
- readOnlyHint: true
- openWorldHint: true
domain_reportDue-diligence report on a website, merchant or counterparty domain: registrable domain, registry RDAP facts (registrar, abuse contact, created, expiry, age, status, nameservers), D…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".
What this check looks forMatched text
inputSchema.properties.domain.description: "...address or URL. Scheme, path, port and credentials are stripped and IDNs are converted to..."
Description, as the server wrote it Due-diligence report on a website, merchant or counterparty domain: registrable domain, registry RDAP facts (registrar, abuse contact, created, expiry, age, status, nameservers), DNS records (A, AAAA, MX, NS, TXT, CAA) from Cloudflare DNS over HTTPS, SPF and DMARC posture, a live web probe (redirect chain, HTTPS, HSTS, site-supplied title, parked or for-sale signs), brand-impersonation checks, and a 0-100 risk score with transparent weighted signals and a verdict. Price: $0.01 per successful call, paid in USDC with x402 v2; failed calls and calls that match nothing are not charged.
Annotations the server declares
- title: Domain due-diligence and risk report
- readOnlyHint: true
- openWorldHint: true
domain_whoisRegistration facts for a domain from its registry's RDAP server, located through the IANA bootstrap file: registrar name, IANA ID and abuse contact, created, updated and expiry dat…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".
What this check looks forMatched text
inputSchema.properties.domain.description: "...address or URL. Scheme, path, port and credentials are stripped and IDNs are converted to..."
Description, as the server wrote it Registration facts for a domain from its registry's RDAP server, located through the IANA bootstrap file: registrar name, IANA ID and abuse contact, created, updated and expiry dates, age and days to expiry, EPP status codes, transfer lock, nameservers, DNSSEC delegation flag, and the registry RDAP URL and terms link. No registrant contact data. Registries whose terms bar reuse (Nominet, CentralNic, Google Registry) are not queried. Price: $0.003 per successful call, paid in USDC with x402 v2; failed calls and calls that match nothing are not charged.
Annotations the server declares
- title: Domain registration facts (RDAP whois)
- readOnlyHint: true
- openWorldHint: true
email_verifyChecks an email address without sending mail: syntax rules (IDN domains in punycode and Unicode), MX records by preference, null MX (RFC 7505), implicit A/AAAA fallback, NXDOMAIN, …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 Checks an email address without sending mail: syntax rules (IDN domains in punycode and Unicode), MX records by preference, null MX (RFC 7505), implicit A/AAAA fallback, NXDOMAIN, SPF and DMARC policy, disposable, role, free-webmail and +tag flags, mail provider from MX hosts, and typo suggestions for popular domains. Returns a verdict and 0-100 score with reasons. Mailbox existence is not confirmed. Sources: Cloudflare DNS over HTTPS, disposable-email-domains (CC0). Price: $0.003 per successful call, paid in USDC with x402 v2; failed calls and calls that match nothing are not charged.
Annotations the server declares
- title: Email address verification (syntax, MX, risk flags)
- readOnlyHint: true
- openWorldHint: true
fx_ratesECB euro foreign exchange reference rates for about 30 currencies, latest or any date since 1999-01-04: base, rate date, rates for the requested currencies, optional converted amou…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 ECB euro foreign exchange reference rates for about 30 currencies, latest or any date since 1999-01-04: base, rate date, rates for the requested currencies, optional converted amounts, and the unmodified ECB euro rates. Rates for non-EUR bases are cross rates calculated by Kit Rx from the ECB euro rates. Daily rates published for information, not transaction quotes. Source: ECB statistics, via the Frankfurter API; the ECB publishes these rates free of charge on its website. Price: $0.002 per successful call, paid in USDC with x402 v2; failed calls and calls that match nothing are not charged.
Annotations the server declares
- title: ECB currency exchange rates and conversion
- readOnlyHint: true
- openWorldHint: true
geo_us_addressGeocodes a one-line US address, or reverse-geocodes a lat/lon point, to coordinates and the Census geographies containing it: state and county FIPS, county subdivision, place, trac…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 Geocodes a one-line US address, or reverse-geocodes a lat/lon point, to coordinates and the Census geographies containing it: state and county FIPS, county subdivision, place, tract, block group, block, ZCTA, congressional and state legislative districts, school districts, urban area, metro or micro area, metro division, CSA, PUMA, census region and division, plus ground elevation in feet. Sources: US Census Geocoder (current benchmark), USGS 3DEP elevation service (EPQS). Price: $0.005 per successful call, paid in USDC with x402 v2; failed calls and calls that match nothing are not charged.
Annotations the server declares
- title: US address geocode with Census geographies
- readOnlyHint: true
- openWorldHint: true
geo_weatherWeather forecast and active alerts for a US address or lat/lon point: NWS office, grid, forecast and county zones, time zone; 7-day periods (temperature, wind, precipitation chance…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 Weather forecast and active alerts for a US address or lat/lon point: NWS office, grid, forecast and county zones, time zone; 7-day periods (temperature, wind, precipitation chance, short and detailed forecast) or, with hourly=true, the next 24 hours adding humidity and dew point; active watches, warnings and advisories with event, severity, urgency, certainty, headline, timing and area. Sources: NOAA National Weather Service API, US Census Geocoder for addresses. Price: $0.005 per successful call, paid in USDC with x402 v2; failed calls and calls that match nothing are not charged.
Annotations the server declares
- title: US weather forecast and active alerts (NWS)
- readOnlyHint: true
- openWorldHint: true
kitrx_catalogFree (no payment).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 Free (no payment). Lists every Kit Rx tool with its price in USD and USDC atomic units, parameters, example arguments and the equivalent HTTP x402 URL, plus the accepted payment networks and payTo address.
Annotations the server declares
- title: Kit Rx catalog
- readOnlyHint: true
- openWorldHint: false
research_briefAI research brief on one question from a live web search: a summary of up to 200 words, up to 8 findings each with citation URLs checked against the search tool's own records (unci…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 AI research brief on one question from a live web search: a summary of up to 200 words, up to 8 findings each with citation URLs checked against the search tool's own records (uncited claims are dropped), the source URLs consulted, and an as-of date. Depth quick (up to 2 searches) or standard (up to 4); optional since date. AI-generated under a fixed Kit Rx prompt and can be wrong. No personal profiles, patient-specific medical advice or investment calls. Price: $0.15 per successful call, paid in USDC with x402 v2; failed calls and calls that match nothing are not charged. AI-generated answer; a call with no arguments returns the stored example answer without a model call.
Annotations the server declares
- title: AI web research brief with cited sources
- readOnlyHint: true
- openWorldHint: true
rx_drugDrug card for a generic or brand name: NLM RxNorm concept (RxCUI, ingredients, brands); one openFDA label (brand checked against RxNorm ingredients, else single-ingredient full pre…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 Drug card for a generic or brand name: NLM RxNorm concept (RxCUI, ingredients, brands); one openFDA label (brand checked against RxNorm ingredients, else single-ingredient full prescribing info, most common route, earliest NDA/BLA) with clipped boxed warning, indications, dosing, contraindications, warnings, interactions, pregnancy, lactation, special-formulation flag, other routes, DailyMed link; DEA schedule and status (NDC data); FDA shortage and 2-year recall summaries. Price: $0.02 per successful call, paid in USDC with x402 v2; failed calls and calls that match nothing are not charged.
Annotations the server declares
- title: Drug card: RxNorm, FDA label, DEA schedule, shortages
- readOnlyHint: true
- openWorldHint: true
rx_label_extractAI extraction from the FDA drug label (openFDA SPL) for an NDC, RxCUI or drug name: boxed warning, indications, dosage forms, renal and hepatic dose adjustment, contraindications, …Reads text others write1 finding
Why it is in this group
- The name "rx_label_extract" says "label": a word for a change.
- 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 server marks "rx_label_extract" read-only (readOnlyHint: true), and its own name, description or input give the class "write".
What this check looks forMatched text
annotations: {"readOnlyHint":true,"openWorldHint":true}; The name "rx_label_extract" says "label": a word for a change.
Description, as the server wrote it AI extraction from the FDA drug label (openFDA SPL) for an NDC, RxCUI or drug name: boxed warning, indications, dosage forms, renal and hepatic dose adjustment, contraindications, pregnancy and lactation, drug interactions, storage. Per section: short summary and points with verbatim label quotes checked in code, plus SPL set ID, version, effective date and DailyMed link. Generated by Anthropic Claude; summarizes FDA labeling, not medical advice. Price: $0.02 per successful call, paid in USDC with x402 v2; failed calls and calls that match nothing are not charged. AI-generated answer; a call with no arguments returns the stored example answer without a model call.
Annotations the server declares
- title: FDA drug label extract: renal, hepatic, boxed warning
- readOnlyHint: true
- openWorldHint: true
rx_ndcFull product record for one package NDC: normalized NDC forms; RxNav status, RxCUI, concept name and history; labeler, marketing category, status and dates, application number, DEA…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 Full product record for one package NDC: normalized NDC forms; RxNav status, RxCUI, concept name and history; labeler, marketing category, status and dates, application number, DEA schedule, imprint, color, shape and size from RxNav; brand and generic name, dosage form, route, active ingredients with strengths and package description from the openFDA NDC Directory; latest CMS NADAC unit cost; openFDA recall enforcement reports from the last 36 months. Price: $0.02 per successful call, paid in USDC with x402 v2; failed calls and calls that match nothing are not charged.
Annotations the server declares
- title: NDC product lookup with price and recalls
- readOnlyHint: true
- openWorldHint: true
rx_ndc_convertConverts a National Drug Code between FDA 10-digit layouts (4-4-2, 5-3-2, 5-4-1) and the HIPAA 11-digit 5-4-2 form, dashed or undashed.Reads text others write1 finding
Why it is in this group
- The name "rx_ndc_convert" says "convert": a word for a change.
- 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 server marks "rx_ndc_convert" read-only (readOnlyHint: true), and its own name, description or input give the class "write".
What this check looks forMatched text
annotations: {"readOnlyHint":true,"openWorldHint":true}; The name "rx_ndc_convert" says "convert": a word for a change.
Description, as the server wrote it Converts a National Drug Code between FDA 10-digit layouts (4-4-2, 5-3-2, 5-4-1) and the HIPAA 11-digit 5-4-2 form, dashed or undashed. Ambiguous undashed 10-digit input is resolved against NLM RxNav by checking all three possible layouts. Returns ndc11, ndc11_dashed, ndc10, product_ndc, segment_format, resolution method, candidate layouts, and RxNav status (ACTIVE, OBSOLETE, ALIEN, UNKNOWN) with RxCUI and concept name. Price: $0.002 per successful call, paid in USDC with x402 v2; failed calls and calls that match nothing are not charged.
Annotations the server declares
- title: NDC format converter (10 to 11 digit)
- readOnlyHint: true
- openWorldHint: true
rx_priceLatest CMS NADAC (National Average Drug Acquisition Cost) per-unit prices for a drug name, brand name or NDC, from data.medicaid.gov (monthly pharmacy invoice survey, weekly files)…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 Latest CMS NADAC (National Average Drug Acquisition Cost) per-unit prices for a drug name, brand name or NDC, from data.medicaid.gov (monthly pharmacy invoice survey, weekly files). Groups by NADAC description, pricing unit (EA, ML, GM), generic or brand rate class and OTC flag, each with min, median and max unit cost across NDCs, effective date, NDC count, sample NDCs, and a 30-unit acquisition cost for per-dose EA items. Single-source generics map to RxNorm brand names. Price: $0.01 per successful call, paid in USDC with x402 v2; failed calls and calls that match nothing are not charged.
Annotations the server declares
- title: Drug acquisition cost (CMS NADAC)
- readOnlyHint: true
- openWorldHint: true
rx_recallsFDA drug recall enforcement reports from openFDA, newest report first, filtered by optional drug, brand or firm text (every word must match; brand names also match their RxNorm ing…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 FDA drug recall enforcement reports from openFDA, newest report first, filtered by optional drug, brand or firm text (every word must match; brand names also match their RxNorm ingredients), look-back months and FDA classification (I, II, III). Entries include recall number, classification, status, product description, reason, recalling firm, distribution pattern, initiation and report dates (ISO) and voluntary or mandated status, plus counts by classification. Price: $0.01 per successful call, paid in USDC with x402 v2; failed calls and calls that match nothing are not charged.
Annotations the server declares
- title: FDA drug recalls (enforcement reports)
- readOnlyHint: true
- openWorldHint: true
rx_shortagesFDA drug shortage records from openFDA, most recently updated first, with optional drug-name and status filters (current, to be discontinued, resolved).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 FDA drug shortage records from openFDA, most recently updated first, with optional drug-name and status filters (current, to be discontinued, resolved). Name filters match every word in generic name, brand or presentation, plus RxNorm ingredients for brand names. Entries give generic name, presentation, dosage form, status, availability, related info, reason, therapeutic category, company, ISO dates and package NDC, plus counts by status and an in_shortage flag. Price: $0.01 per successful call, paid in USDC with x402 v2; failed calls and calls that match nothing are not charged.
Annotations the server declares
- title: FDA drug shortages
- readOnlyHint: true
- openWorldHint: true
sec_digestAI-generated digest of one SEC EDGAR filing section: 10-K/10-Q risk factors (Item 1A) or MD&A, or 8-K item summaries with press-release exhibits.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 AI-generated digest of one SEC EDGAR filing section: 10-K/10-Q risk factors (Item 1A) or MD&A, or 8-K item summaries with press-release exhibits. By accession, or ticker/CIK plus form (latest filing). Key points each carry a verbatim quote checked against the filing text and an EDGAR URL; optional new and removed risk factors versus the prior same-form filing. Generated by Anthropic Claude under a fixed Kit Rx prompt; not investment advice. Price: $0.02 per successful call, paid in USDC with x402 v2; failed calls and calls that match nothing are not charged. AI-generated answer; a call with no arguments returns the stored example answer without a model call.
Annotations the server declares
- title: AI digest of SEC 10-K, 10-Q and 8-K filings
- readOnlyHint: true
- openWorldHint: true
sec_eventsRecent Form 8-K current reports (material corporate events) for a ticker or CIK over the last 1 to 365 days: filing date, event date, accession number, document URL and item number…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 Recent Form 8-K current reports (material corporate events) for a ticker or CIK over the last 1 to 365 days: filing date, event date, accession number, document URL and item numbers decoded to their Form 8-K titles (1.01 material agreement, 2.02 earnings results, 5.02 director or officer changes, 7.01 Reg FD, 8.01 other events and more), plus counts by item. Source: SEC EDGAR submissions API. Price: $0.005 per successful call, paid in USDC with x402 v2; failed calls and calls that match nothing are not charged.
Annotations the server declares
- title: Recent 8-K material events from SEC EDGAR
- readOnlyHint: true
- openWorldHint: true
sec_filingsSEC EDGAR profile and latest filings for a ticker or CIK: company name, CIK, SIC code and industry, fiscal year end, state of incorporation, tickers and exchanges, plus the newest …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 SEC EDGAR profile and latest filings for a ticker or CIK: company name, CIK, SIC code and industry, fiscal year end, state of incorporation, tickers and exchanges, plus the newest filings of selected forms (10-K, 10-Q, 8-K, 20-F, 6-K, S-1, DEF 14A, 4, 13F-HR, SC 13G) with filing and report dates, accession number, 8-K items and primary document URL. Optional latest XBRL revenue, net income, total assets and cash. Sources: SEC EDGAR submissions and XBRL APIs. Price: $0.01 per successful call, paid in USDC with x402 v2; failed calls and calls that match nothing are not charged.
Annotations the server declares
- title: Company filings and key financials from SEC EDGAR
- readOnlyHint: true
- openWorldHint: true
token_priceSpot USD and native-coin price of an ERC-20 token on Base, Ethereum, Arbitrum, OP Mainnet, Polygon or Avalanche from its deepest Uniswap v2/v3 pool (plus Aerodrome on Base) paired …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.examples[0] contains a 42-character base64 string.
What this check looks forMatched text
inputSchema.properties.address.examples[0]: "0x940181a94A35A4569E4529A3CDfB74e38FD98631"
Description, as the server wrote it Spot USD and native-coin price of an ERC-20 token on Base, Ethereum, Arbitrum, OP Mainnet, Polygon or Avalanche from its deepest Uniswap v2/v3 pool (plus Aerodrome on Base) paired with the wrapped native coin or USDC, with each pool's fee and USD liquidity, the best pool, a confidence level and the block read. Sources: on-chain reads via PublicNode JSON-RPC (official RPC fallback); native USD from Uniswap v3 pools. Price: $0.002 per successful call, paid in USDC with x402 v2; failed calls and calls that match nothing are not charged.
Annotations the server declares
- title: ERC-20 token USD price from on-chain DEX pools
- readOnlyHint: true
- openWorldHint: true
token_scanHeuristic risk screen of an ERC-20 token on Base, Ethereum, Arbitrum, OP Mainnet, Polygon or Avalanche: name, symbol, supply; owner and renounce status; proxy and upgradeability; S…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.examples[0] contains a 42-character base64 string.
What this check looks forMatched text
inputSchema.properties.address.examples[0]: "0x4ed4E862860beD51a9570b96d89aF5E1B0Efefed"
Description, as the server wrote it Heuristic risk screen of an ERC-20 token on Base, Ethereum, Arbitrum, OP Mainnet, Polygon or Avalanche: name, symbol, supply; owner and renounce status; proxy and upgradeability; Sourcify verification; privileged functions in bytecode or verified ABI (mint, pause, blacklist, fee, limit, trading switch); Uniswap v2/v3 and Aerodrome pools with USD depth; risk score 0-100 with a reason per flag. Sources: on-chain reads via PublicNode JSON-RPC, Sourcify. Price: $0.01 per successful call, paid in USDC with x402 v2; failed calls and calls that match nothing are not charged.
Annotations the server declares
- title: ERC-20 token risk scan (owner, proxy, liquidity)
- readOnlyHint: true
- openWorldHint: true
tx_decodeDecoded EVM transaction on Base, Ethereum, Arbitrum, OP Mainnet, Polygon or Avalanche: status, block time, sender, target, value and fee in native coin and USD (current price), met…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 Decoded EVM transaction on Base, Ethereum, Arbitrum, OP Mainnet, Polygon or Avalanche: status, block time, sender, target, value and fee in native coin and USD (current price), method and arguments, decoded events, ERC-20/721 token transfers, net token flows per address and a one-line summary. ABIs from Sourcify verified contracts or the Sourcify 4byte database. Ethereum receipts cover about the last 2 days. Sources: PublicNode and official chain JSON-RPC, Uniswap v3 pools. Price: $0.005 per successful call, paid in USDC with x402 v2; failed calls and calls that match nothing are not charged.
Annotations the server declares
- title: Decode an EVM transaction: method, events, transfers
- readOnlyHint: true
- openWorldHint: true
wallet_balancesBalances of an EVM wallet (0x address, ENS name or Basename) on Base, Ethereum, Arbitrum, OP Mainnet, Polygon and Avalanche: native coin plus a curated list of 9-13 major tokens pe…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 Balances of an EVM wallet (0x address, ENS name or Basename) on Base, Ethereum, Arbitrum, OP Mainnet, Polygon and Avalanche: native coin plus a curated list of 9-13 major tokens per chain (USDC, USDT, DAI, WETH, WBTC/cbBTC, AERO, ARB, OP and more), USD values, per-chain and total USD, chains holding funds, and native USDC per chain for x402 payments. Sources: each chain's JSON-RPC via PublicNode; prices from on-chain Uniswap v3 and Velodrome pools. Price: $0.003 per successful call, paid in USDC with x402 v2; failed calls and calls that match nothing are not charged.
Annotations the server declares
- title: Multi-chain wallet balances in USD (curated tokens)
- readOnlyHint: true
- openWorldHint: true
web_answerAI question answering over 1 to 3 web page, PDF or DOCX URLs: grounded, cited RAG answers for research or fact checks.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.urls): a request to a URL can carry data to whoever runs it.
Findings
- Info
The input schema at properties.urls.description mentions "credentials".
What this check looks forMatched text
inputSchema.properties.urls.description: "..., ODT, plain text). Default ports only; credentials, IP addresses, localhost and private or..."
Description, as the server wrote it AI question answering over 1 to 3 web page, PDF or DOCX URLs: grounded, cited RAG answers for research or fact checks. Pages are fetched with SSRF checks and read as Markdown (about the first 8k tokens each). Returns an answer of at most 150 words, claims with source URL and verbatim quote (checked in code against the page; unverified claims dropped) and confidence. 404, not charged, when the pages do not support an answer. AI-generated by Anthropic Claude. Price: $0.01 per successful call, paid in USDC with x402 v2; failed calls and calls that match nothing are not charged. AI-generated answer; a call with no arguments returns the stored example answer without a model call.
Annotations the server declares
- title: AI answer from web pages with verified quotes
- readOnlyHint: true
- openWorldHint: true
web_checkFacts about a URL from a live fetch, never its body: final URL, status and chain after up to 5 validated redirects, response time, content type, server, security headers present/mi…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".
What this check looks forMatched text
inputSchema.properties.url.description: "...ecked as https://). Default ports only; credentials, IP addresses, localhost and private or..."
Description, as the server wrote it Facts about a URL from a live fetch, never its body: final URL, status and chain after up to 5 validated redirects, response time, content type, server, security headers present/missing (HSTS, CSP, nosniff, framing, Referrer-Policy, Permissions-Policy), title, description, canonical, lang, Open Graph, robots meta, robots.txt rules for GPTBot, ClaudeBot, Google-Extended, CCBot, PerplexityBot, *, llms.txt, agent card, OpenAPI, x402 discovery, 402 terms. DNS via Cloudflare DoH. Price: $0.005 per successful call, paid in USDC with x402 v2; failed calls and calls that match nothing are not charged.
Annotations the server declares
- title: Website check: redirects, headers, SEO, agent readiness
- readOnlyHint: true
- openWorldHint: true
web_extractAI web scraping to structured JSON: the fields of a JSON Schema (object root, up to 4 KB) or a field-name list, extracted from one public web page or document (HTML, PDF, DOCX) fet…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".
What this check looks forMatched text
inputSchema.properties.url.description: "...DOCX, spreadsheet). Default ports only; credentials, IP addresses, localhost and private or..."
Description, as the server wrote it AI web scraping to structured JSON: the fields of a JSON Schema (object root, up to 4 KB) or a field-name list, extracted from one public web page or document (HTML, PDF, DOCX) fetched via SSRF-checked redirects. Returns data in the requested shape, null where the page states no value, a page quote per value when one matches the page text in code, final URL, title and truncation flag. AI-generated by Anthropic Claude under a fixed Kit Rx prompt; values can be wrong. Price: $0.005 per successful call, paid in USDC with x402 v2; failed calls and calls that match nothing are not charged. AI-generated answer; a call with no arguments returns the stored example answer without a model call.
Annotations the server declares
- title: AI structured data extraction from a web page
- readOnlyHint: true
- openWorldHint: true
web_headersGrades a URL's HTTP security headers A+ to F (0-100) from one live fetch after up to 5 validated redirects.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".
What this check looks forMatched text
inputSchema.properties.url.description: "...ecked as https://). Default ports only; credentials, IP addresses, localhost and private or..."
Description, as the server wrote it Grades a URL's HTTP security headers A+ to F (0-100) from one live fetch after up to 5 validated redirects. Per header: present, value, rating, points, issue, fix for HSTS, Content-Security-Policy (unsafe-inline, unsafe-eval, wildcards, frame-ancestors), X-Content-Type-Options, X-Frame-Options, Referrer-Policy, Permissions-Policy, COOP, COEP, CORP; Server/X-Powered-By version leaks; cookie Secure, HttpOnly, SameSite flags; missing list, summary. DNS via Cloudflare DoH. Price: $0.005 per successful call, paid in USDC with x402 v2; failed calls and calls that match nothing are not charged.
Annotations the server declares
- title: HTTP Security Headers Check (A+ to F grade)
- readOnlyHint: true
- openWorldHint: true
web_markdownMarkdown of a live web page for LLM context: final URL after up to 5 validated redirects, HTTP status, content type, title, meta description, lang, Markdown (main content by defaul…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".
What this check looks forMatched text
inputSchema.properties.url.description: "...tched as https://). Default ports only; credentials, IP addresses, localhost and private or..."
Description, as the server wrote it Markdown of a live web page for LLM context: final URL after up to 5 validated redirects, HTTP status, content type, title, meta description, lang, Markdown (main content by default: <main>/<article>, without scripts, navigation, asides, footers), word count, token estimate, link count and truncation flag. PDF, DOCX, XLSX, OpenDocument, CSV and XML URLs are converted too; plain text is passed through. Conversion: Cloudflare Workers AI toMarkdown. DNS via Cloudflare DoH. Price: $0.003 per successful call, paid in USDC with x402 v2; failed calls and calls that match nothing are not charged.
Annotations the server declares
- title: Web page to clean Markdown
- readOnlyHint: true
- openWorldHint: true
x402_receiptDecodes the x402 v2 PAYMENT-SIGNATURE header sent with the request into a readable receipt: x402 version, scheme, CAIP-2 network and chain name, resource URL, the copied payment re…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 Decodes the x402 v2 PAYMENT-SIGNATURE header sent with the request into a readable receipt: x402 version, scheme, CAIP-2 network and chain name, resource URL, the copied payment requirements, payer, recipient, value in atomic units and USD, USDC contract, nonce, validity window in ISO-8601 with seconds to expiry, signature length and prefix, extension keys, and pass/fail diagnostics. Computed locally, no upstream source. Price: $0.002 per successful call, paid in USDC with x402 v2; failed calls and calls that match nothing are not charged.
Annotations the server declares
- title: x402 payment receipt: decode the payment you sent
- readOnlyHint: true
- openWorldHint: true
x402_recommendAI recommendation of the best paid x402 API for a task in plain words.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 AI recommendation of the best paid x402 API for a task in plain words. Ultravioleta Bazaar and 402 Index listings are deduped and keyword-scored; Anthropic Claude reranks up to 30 for fit. The first 3 GET picks get a live unpaid preflight: valid 402, live USDC price, networks, risk flags, OFAC payTo screen, domain age. Each result: URL, description, price, networks, usage evidence, preflight, AI-written reason, verbatim-checked listing quote. Kit Rx routes are labelled. Price: $0.005 per successful call, paid in USDC with x402 v2; failed calls and calls that match nothing are not charged. AI-generated answer; a call with no arguments returns the stored example answer without a model call.
Annotations the server declares
- title: AI x402 API recommender: ranked, preflighted picks
- readOnlyHint: true
- openWorldHint: true
Only reads 6Show 6 read-only tools
chain_sign_inspectInspects EIP-712 typed data before signing: kind (EIP-2612 or DAI permit, Permit2 allowance or transfer, EIP-3009 transfer authorization as used by x402, Seaport order, Safe transa…Reads text others write1 finding
Why it is in this group
- The name "chain_sign_inspect" says "inspect": a word for reading.
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.expected_spender.examples[0] contains a 42-character base64 string.
What this check looks forMatched text
inputSchema.properties.expected_spender.examples[0]: "0xAf8E0A8a8A0A8f65831b2dD0FFe0c1B804698D70"
Description, as the server wrote it Inspects EIP-712 typed data before signing: kind (EIP-2612 or DAI permit, Permit2 allowance or transfer, EIP-3009 transfer authorization as used by x402, Seaport order, Safe transaction, generic), digest, chain id with mismatch check, verifying contract with token symbol and decimals, spender or recipient, raw and formatted amount, unlimited flag, deadline window in ISO time, risk level, flags with reasons, one-line summary. Sources: JSON-RPC via PublicNode, Sourcify. Price: $0.002 per successful call, paid in USDC with x402 v2; failed calls and calls that match nothing are not charged.
Annotations the server declares
- title: EIP-712 signature request inspector (permits, x402)
- readOnlyHint: true
- openWorldHint: true
ens_resolveENS and Basenames lookup: name to address, or address to primary name with forward verification.Reads text others write1 finding
Why it is in this group
- The name "ens_resolve" says "resolve": a word for reading.
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.examples[0] contains a 42-character base64 string.
What this check looks forMatched text
inputSchema.properties.address.examples[0]: "0xd8dA6BF26964aF9D7eEd9e03E53415D37aA96045"
Description, as the server wrote it ENS and Basenames lookup: name to address, or address to primary name with forward verification. Returns the ETH address, resolver contract, text records (url, avatar, description, Twitter, GitHub, Telegram, and email when set on-chain) and decoded contenthash (IPFS, IPNS, Swarm, Arweave). Sources: ENS Universal Resolver on Ethereum and the Basenames registry on Base, read via PublicNode JSON-RPC; offchain (CCIP-read) names are reported, not followed. Price: $0.002 per successful call, paid in USDC with x402 v2; failed calls and calls that match nothing are not charged.
Annotations the server declares
- title: ENS and Basenames resolver (forward and reverse)
- readOnlyHint: true
- openWorldHint: true
sec_searchFull-text search of SEC EDGAR filings since 2001 (10-K, 10-Q, 8-K, proxy, S-1, exhibits and more): keywords, "exact phrase", OR, NOT, NEAR and trailing * wildcard, filtered by form…Reads text others writeNo findings
Why it is in this group
- The name "sec_search" says "search": a word for reading.
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 Full-text search of SEC EDGAR filings since 2001 (10-K, 10-Q, 8-K, proxy, S-1, exhibits and more): keywords, "exact phrase", OR, NOT, NEAR and trailing * wildcard, filtered by form type, ticker or CIK and filing date range. Returns total hits, hits by form, top filers and matching documents with company, CIK, ticker, form, filing date, period, accession number, 8-K items and sec.gov document URL. Insider Forms 3, 4, 5 and 144 excluded. Source: SEC EDGAR Full-Text Search. Price: $0.005 per successful call, paid in USDC with x402 v2; failed calls and calls that match nothing are not charged.
Annotations the server declares
- title: SEC EDGAR full-text search of company filings
- readOnlyHint: true
- openWorldHint: true
text_summarizeAI summary of up to 8,000 characters of text, or of a public web page URL (main content fetched through SSRF-validated redirects and converted to Markdown, first 30,000 characters)…Reads text others writeCan send data out1 finding
Why it is in this group
- The name "text_summarize" says "summarize": a word for reading.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
- untrusted-inputThe description's first sentence says "web": it reads web pages, which anyone can write.
- 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.
Findings
- Info
The input schema at properties.url.description mentions "credentials".
What this check looks forMatched text
inputSchema.properties.url.description: "...ze instead of text. Default ports only; credentials, IP addresses, localhost and private or..."
Description, as the server wrote it AI summary of up to 8,000 characters of text, or of a public web page URL (main content fetched through SSRF-validated redirects and converted to Markdown, first 30,000 characters). Length short, medium or long; format paragraph, bullets or json. Returns summary, key points, detected language, source URL and title, input size and truncation flag. Generated by Anthropic Claude under a fixed Kit Rx prompt; AI-generated text may omit or misstate details. Price: $0.006 per successful call, paid in USDC with x402 v2; failed calls and calls that match nothing are not charged. AI-generated answer; a call with no arguments returns the stored example answer without a model call.
Annotations the server declares
- title: AI summary of text or a web page
- readOnlyHint: true
- openWorldHint: true
web_readAI question answering over one live web page or document URL: an answer drawn only from the page text, verbatim quotes with character offsets checked against the page, an answerabl…Reads private dataReads text others writeCan send data out1 finding
Why it is in this group
- The name "web_read" says "read": a word for reading.
What it can reach
- private-dataThe description's first sentence says "document": it reads documents.
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
- untrusted-inputThe name "web_read" says "read" and "web": it reads web pages, which anyone can write.
- untrusted-inputThe description's first sentence says "web": it reads web pages, which anyone can write.
- 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.
Findings
- Info
The input schema at properties.url.description mentions "credentials".
What this check looks forMatched text
inputSchema.properties.url.description: "...tched as https://). Default ports only; credentials, IP addresses, localhost and private or..."
Description, as the server wrote it AI question answering over one live web page or document URL: an answer drawn only from the page text, verbatim quotes with character offsets checked against the page, an answerable flag and low/medium/high confidence. answerable:false (page lacks the answer) is a normal, charged 200 result. Without a question: summary, key points, title. Also final URL, page size, truncation flag. Generated by Anthropic Claude under a fixed Kit Rx prompt; may miss or misstate details. Price: $0.004 per successful call, paid in USDC with x402 v2; failed calls and calls that match nothing are not charged. AI-generated answer; a call with no arguments returns the stored example answer without a model call.
Annotations the server declares
- title: Ask a web page a question, with verified quotes
- readOnlyHint: true
- openWorldHint: true
x402_findSearch and discovery of paid x402 API endpoints for a plain-language need, across Ultravioleta DAO x402 Bazaar and 402 Index directories.Reads private dataReads text others writeNo findings
Why it is in this group
- The name "x402_find" says "find": a word for reading.
What it can reach
- private-dataThe description's first sentence says "search" and "directory": it reads files.
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
Description, as the server wrote it Search and discovery of paid x402 API endpoints for a plain-language need, across Ultravioleta DAO x402 Bazaar and 402 Index directories. Dedupes by resource URL; ranks by relevance, verification, uptime, price and recency. Each result: resource URL, short description, USDC price, CAIP-2 networks, sources, verification. The top 3 GET endpoints get a live unpaid check: valid x402 402, live price, networks, canonical USDC, risk flags, OFAC SDN payTo screen, RDAP domain age. Price: $0.01 per successful call, paid in USDC with x402 v2; failed calls and calls that match nothing are not charged.
Annotations the server declares
- title: x402 finder: search paid APIs, ranked and live-checked
- readOnlyHint: true
- openWorldHint: true
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.
1152 paths in this server alone from private data, through text someone else wrote, to a way out. Denying 2 private-data tools breaks every one.
- Reads private data2
- Reads files, mail, messages, databases, secrets, calendars or repositories.
- Reads text others write48
- Brings in text someone else can write: web pages, inbound email, messages, issues and comments.
- Can send data out12
- Can carry data to people or systems: it sends, posts or publishes, or takes a recipient, channel or URL.
{
"permissions": {
"deny": [
"mcp__kitrx__web_read",
"mcp__kitrx__x402_find"
]
}
}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 15, 2026, 13:18 UTC, unchanged since Nov 15, 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
- 32
- 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 32 tools that change things (2 irreversible, 30 write). Whether a call also works without credentials was not tested: tools are never called.
What this check looks forMatched text
https://api.kit-rx.com/mcp: tools/list returned 36 tools to an anonymous client, including x402_preflight (irreversible), chain_calldata_decode (irreversible), rx_drug (write), rx_shortages (write), rx_recalls (write) and 27 more
Findings
Matches of fixed checks, with the text they matched.
- Warnrx_ndc_convert
The server marks "rx_ndc_convert" read-only (readOnlyHint: true), and its own name, description or input give the class "write".
What this check looks forMatched text
annotations: {"readOnlyHint":true,"openWorldHint":true}; The name "rx_ndc_convert" says "convert": a word for a change.
- Warnx402_preflight
The server marks "x402_preflight" read-only (readOnlyHint: true), and its own name, description or input give the class "irreversible".
What this check looks forMatched text
annotations: {"readOnlyHint":true,"openWorldHint":true}; The description's first verb is "send": a word for sending, and what is sent cannot be taken back.
The input schema at properties.address.examples[0] contains a 42-character base64 string.
What this check looks forMatched text
inputSchema.properties.address.examples[0]: "0x098B716B8Aaf21512996dC57EB0615e2383E2f96"
The input schema at properties.to.examples[0] contains a 42-character base64 string.
What this check looks forMatched text
inputSchema.properties.to.examples[0]: "0x833589fCD6eDb6E08f4c7C32D4f71b54bdA02913"
The server marks "chain_calldata_decode" read-only (readOnlyHint: true), and its own name, description or input give the class "irreversible".
What this check looks forMatched text
annotations: {"readOnlyHint":true,"openWorldHint":true}; The input addresses people (input.to): what it sends cannot be taken back.
The input schema at properties.expected_spender.examples[0] contains a 42-character base64 string.
What this check looks forMatched text
inputSchema.properties.expected_spender.examples[0]: "0xAf8E0A8a8A0A8f65831b2dD0FFe0c1B804698D70"
Show 20 more findingsShow fewer
- Warncontract_abi
The input schema at properties.address.examples[0] contains a 42-character base64 string.
What this check looks forMatched text
inputSchema.properties.address.examples[0]: "0x833589fCD6eDb6E08f4c7C32D4f71b54bdA02913"
- Warntoken_scan
The input schema at properties.address.examples[0] contains a 42-character base64 string.
What this check looks forMatched text
inputSchema.properties.address.examples[0]: "0x4ed4E862860beD51a9570b96d89aF5E1B0Efefed"
- Warntoken_price
The input schema at properties.address.examples[0] contains a 42-character base64 string.
What this check looks forMatched text
inputSchema.properties.address.examples[0]: "0x940181a94A35A4569E4529A3CDfB74e38FD98631"
- Warnens_resolve
The input schema at properties.address.examples[0] contains a 42-character base64 string.
What this check looks forMatched text
inputSchema.properties.address.examples[0]: "0xd8dA6BF26964aF9D7eEd9e03E53415D37aA96045"
- Warntext_translate
The server marks "text_translate" read-only (readOnlyHint: true), and its own name, description or input give the class "irreversible".
What this check looks forMatched text
annotations: {"readOnlyHint":true,"openWorldHint":true}; The input addresses people (input.to): what it sends cannot be taken back.
- Warnrx_label_extract
The server marks "rx_label_extract" read-only (readOnlyHint: true), and its own name, description or input give the class "write".
What this check looks forMatched text
annotations: {"readOnlyHint":true,"openWorldHint":true}; The name "rx_label_extract" says "label": a word for a change.
- Infodomain_report
The input schema at properties.domain.description mentions "credentials".
What this check looks forMatched text
inputSchema.properties.domain.description: "...address or URL. Scheme, path, port and credentials are stripped and IDNs are converted to..."
- Infodomain_whois
The input schema at properties.domain.description mentions "credentials".
What this check looks forMatched text
inputSchema.properties.domain.description: "...address or URL. Scheme, path, port and credentials are stripped and IDNs are converted to..."
- Infodomain_dns
The input schema at properties.domain.description mentions "credentials".
What this check looks forMatched text
inputSchema.properties.domain.description: "...address or URL. Scheme, path, port and credentials are stripped and IDNs are converted to..."
The input schema at properties.domain.description mentions "credentials".
What this check looks forMatched text
inputSchema.properties.domain.description: "...address or URL. Scheme, path, port and credentials are stripped and IDNs are converted to..."
- Infoweb_check
The input schema at properties.url.description mentions "credentials".
What this check looks forMatched text
inputSchema.properties.url.description: "...ecked as https://). Default ports only; credentials, IP addresses, localhost and private or..."
- Infoweb_headers
The input schema at properties.url.description mentions "credentials".
What this check looks forMatched text
inputSchema.properties.url.description: "...ecked as https://). Default ports only; credentials, IP addresses, localhost and private or..."
- Infoweb_markdown
The input schema at properties.url.description mentions "credentials".
What this check looks forMatched text
inputSchema.properties.url.description: "...tched as https://). Default ports only; credentials, IP addresses, localhost and private or..."
- Infodoc_markdown
The input schema at properties.url.description mentions "credentials".
What this check looks forMatched text
inputSchema.properties.url.description: "...le (at most 10 MB). Default ports only; credentials, IP addresses, localhost and private or..."
- Infox402_preflight
The input schema at properties.url.description mentions "credentials".
What this check looks forMatched text
inputSchema.properties.url.description: "...racters. Default port only; no embedded credentials, IP-address hosts, localhost or private..."
- Infotext_summarize
The input schema at properties.url.description mentions "credentials".
What this check looks forMatched text
inputSchema.properties.url.description: "...ze instead of text. Default ports only; credentials, IP addresses, localhost and private or..."
- Infoweb_read
The input schema at properties.url.description mentions "credentials".
What this check looks forMatched text
inputSchema.properties.url.description: "...tched as https://). Default ports only; credentials, IP addresses, localhost and private or..."
- Infoweb_extract
The input schema at properties.url.description mentions "credentials".
What this check looks forMatched text
inputSchema.properties.url.description: "...DOCX, spreadsheet). Default ports only; credentials, IP addresses, localhost and private or..."
- Infoweb_answer
The input schema at properties.urls.description mentions "credentials".
What this check looks forMatched text
inputSchema.properties.urls.description: "..., ODT, plain text). Default ports only; credentials, IP addresses, localhost and private or..."
- Infodoc_ocr
The input schema at properties.url.description mentions "credentials".
What this check looks forMatched text
inputSchema.properties.url.description: "...nd 8000 x 8000 px). Default ports only; credentials, IP addresses, localhost and private or..."
History
Changes to its tool definitions and to its remote endpoint, newest first.
Tools13 added.+13 addedView diff
- Added:company_briefChanges things
- Added:company_brief_proChanges things
- Added:company_reportChanges things
- Added:doc_ocrChanges things
- Added:research_briefChanges things
- Added:rx_label_extractChanges things
- Added:sec_digestChanges things
- Added:text_summarizeOnly reads
- Added:text_translateCan't be undone
- Added:web_answerChanges things
- Added:web_extractChanges things
- Added:web_readOnly reads
- Added:x402_recommendChanges things
Tools17 added, 9 changed.+17 added~9 changedView diff
- Added:chain_calldata_decodeCan't be undone
- Added:chain_gasChanges things
- Added:chain_sign_inspectOnly reads
- Added:compliance_sanctionsChanges things
- Added:contract_abiChanges things
- Added:doc_markdownChanges things
- Added:ens_resolveOnly reads
- Added:fx_ratesChanges things
- Added:sec_eventsChanges things
- Added:sec_searchOnly reads
- Added:token_priceChanges things
- Added:token_scanChanges things
- Added:tx_decodeChanges things
- Added:wallet_balancesChanges things
- Added:web_headersChanges things
- Added:web_markdownChanges things
- Added:x402_findOnly reads
- Changed:email_verifyChanges thingsdescription
- Changed:geo_us_addressChanges thingsinputSchema
- Changed:geo_weatherChanges thingsinputSchema
- Changed:rx_ndcChanges thingsdescription
- Changed:rx_priceChanges thingsinputSchema
- Changed:rx_recallsChanges thingsdescription
- Changed:rx_shortagesChanges thingsdescription
- Changed:x402_preflightCan't be undonedescription
- Changed:x402_receiptChanges thingsdescription
Install
Pinned to version 1.3.0. Values in angle brackets are yours to fill in.
claude mcp add --transport http kitrx https://api.kit-rx.com/mcpclaude mcp add --transport http kitrx https://api.kit-rx.com/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__kitrx__sec_search",
"mcp__kitrx__x402_find",
"mcp__kitrx__chain_sign_inspect",
"mcp__kitrx__ens_resolve"
],
"ask": [
"mcp__kitrx__rx_drug",
"mcp__kitrx__rx_shortages",
"mcp__kitrx__rx_recalls",
"mcp__kitrx__rx_ndc_convert",
"mcp__kitrx__rx_ndc",
"mcp__kitrx__rx_price",
"mcp__kitrx__domain_report",
"mcp__kitrx__domain_whois",
"mcp__kitrx__domain_dns",
"mcp__kitrx__domain_email_auth",
"mcp__kitrx__company_profile",
"mcp__kitrx__sec_filings",
"mcp__kitrx__sec_events",
"mcp__kitrx__fx_rates",
"mcp__kitrx__web_check",
"mcp__kitrx__web_headers",
"mcp__kitrx__web_markdown",
"mcp__kitrx__doc_markdown",
"mcp__kitrx__geo_us_address",
"mcp__kitrx__geo_weather",
"mcp__kitrx__email_verify",
"mcp__kitrx__x402_receipt",
"mcp__kitrx__x402_preflight",
"mcp__kitrx__compliance_sanctions",
"mcp__kitrx__chain_gas",
"mcp__kitrx__chain_calldata_decode",
"mcp__kitrx__contract_abi",
"mcp__kitrx__tx_decode",
"mcp__kitrx__token_scan",
"mcp__kitrx__token_price",
"mcp__kitrx__wallet_balances",
"mcp__kitrx__text_translate",
"mcp__kitrx__text_summarize",
"mcp__kitrx__web_read",
"mcp__kitrx__web_extract",
"mcp__kitrx__company_brief",
"mcp__kitrx__x402_recommend",
"mcp__kitrx__web_answer",
"mcp__kitrx__rx_label_extract",
"mcp__kitrx__sec_digest",
"mcp__kitrx__doc_ocr",
"mcp__kitrx__research_brief",
"mcp__kitrx__company_brief_pro",
"mcp__kitrx__company_report",
"mcp__kitrx__kitrx_catalog"
],
"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 kitrx. Review it before you use it.
Capability tags behind 48 rules
- mcp__kitrx__rx_drug untrusted-input
- mcp__kitrx__rx_shortages untrusted-input
- mcp__kitrx__rx_recalls untrusted-input
- mcp__kitrx__rx_ndc_convert untrusted-input
- mcp__kitrx__rx_ndc untrusted-input
- mcp__kitrx__rx_price untrusted-input
- mcp__kitrx__domain_report untrusted-input
- mcp__kitrx__domain_whois untrusted-input
- mcp__kitrx__domain_dns untrusted-input
- mcp__kitrx__domain_email_auth untrusted-input
- mcp__kitrx__company_profile untrusted-input
- mcp__kitrx__sec_filings untrusted-input
- mcp__kitrx__sec_search untrusted-input
- mcp__kitrx__sec_events untrusted-input
- mcp__kitrx__fx_rates untrusted-input
- mcp__kitrx__web_check untrusted-input, outbound
- mcp__kitrx__web_headers untrusted-input, outbound
- mcp__kitrx__web_markdown untrusted-input, outbound
- mcp__kitrx__doc_markdown untrusted-input, outbound
- mcp__kitrx__geo_us_address untrusted-input
- mcp__kitrx__geo_weather untrusted-input
- mcp__kitrx__email_verify untrusted-input
- mcp__kitrx__x402_receipt untrusted-input
- mcp__kitrx__x402_preflight untrusted-input, outbound
- mcp__kitrx__x402_find private-data, untrusted-input
- mcp__kitrx__compliance_sanctions untrusted-input
- mcp__kitrx__chain_gas untrusted-input
- mcp__kitrx__chain_calldata_decode untrusted-input, outbound
- mcp__kitrx__chain_sign_inspect untrusted-input
- mcp__kitrx__contract_abi untrusted-input
- mcp__kitrx__tx_decode untrusted-input
- mcp__kitrx__token_scan untrusted-input
- mcp__kitrx__token_price untrusted-input
- mcp__kitrx__wallet_balances untrusted-input
- mcp__kitrx__ens_resolve untrusted-input
- mcp__kitrx__text_translate untrusted-input, outbound
- mcp__kitrx__text_summarize untrusted-input, outbound
- mcp__kitrx__web_read private-data, untrusted-input, outbound
- mcp__kitrx__web_extract untrusted-input, outbound
- mcp__kitrx__company_brief untrusted-input
- mcp__kitrx__x402_recommend untrusted-input
- mcp__kitrx__web_answer untrusted-input, outbound
- mcp__kitrx__rx_label_extract untrusted-input
- mcp__kitrx__sec_digest untrusted-input
- mcp__kitrx__doc_ocr untrusted-input, outbound
- mcp__kitrx__research_brief untrusted-input
- mcp__kitrx__company_brief_pro untrusted-input
- mcp__kitrx__company_report 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/com.kit-rx.api/kitrx/verify?hash=59729d8049a41e33df4f1eaec61ab0891ec60224b8325acf0d9df4ad02b4defe" | jq -e '.changed == false'Pinned hash 59729d8049a4, from Nov 16, 2026, 00:18 UTC.
Watch
Get told when its tools change: a new tool, a new description, a new class.
https://radar.anlyon.com/servers/com.kit-rx.api/kitrx/changes.xmlAn Atom feed. Works in any feed reader or Slack's RSS app.
Badge
[](https://radar.anlyon.com/servers/com.kit-rx.api/kitrx)Source: MCP Registry. Synced Nov 16, 2026, 00:17 UTC.
Kit Rx MCP server: questions and answers
Answered from the scan above, not from the publisher's description.
Is the Kit Rx MCP server safe?
- No scan can promise that. What its definitions show: 49 tools, 43 of which can change, send or delete something, and 26 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 Kit Rx MCP server delete data?
- No tool is classed as destructive: going by their definitions, none of them deletes, cancels or revokes anything.
Can the Kit Rx MCP server do things that cannot be undone?
- Yes. 3 tools can send, pay, publish or merge: x402_preflight, chain_calldata_decode and text_translate.
What tools does the Kit Rx MCP server have?
- 49 tools: 3 irreversible, 40 write, 6 read. x402_preflight, chain_calldata_decode, text_translate, rx_drug, rx_shortages, rx_recalls, rx_ndc_convert, rx_ndc and 41 more.
Does the Kit Rx MCP server need sign-in or OAuth?
- Its remote endpoint listed its tools to an anonymous client, without sign-in.
Does the Kit Rx MCP server send data to outside services?
- 12 tools can carry data out (it sends, posts or publishes, or takes a recipient, channel or URL): web_check, web_headers, web_markdown, doc_markdown, x402_preflight, chain_calldata_decode and 6 more.
Has the Kit Rx MCP server changed its tools?
- Yes, 2 times since Anlyon Radar started tracking it, most recently on Nov 16, 2026. Every change is on this page and in its Atom feed.