x402 Verification API

Honeypot and rug-pull safety checker for ERC20 tokens on Base. Analyzes token contracts for common scam patterns and risky functions.

IBAN validation and BIC/SWIFT lookup backed by primary bank registries (GLEIF, SIX, Bundesbank, EBA, Bank of England). Returns bank identity, SEPA and Verification of Payee reachability, sanctions and risk flags, Swiss clearing detail, ISO 20022 structured addresses. Pay per call in USDC via x402, no account needed. MCP server available for AI agents.

Pre-send risk check for AI agents on Base. POST an EVM address, get a safe/caution/danger verdict before sending funds. Signals: OFAC sanctions, scam and exploit lists, live address-poisoning detection (lookalike and fake-token scammers), contract deployer and verification checks, on-chain age/activity, and a permanent reputation graph. Unknown data never returns safe. Tiers quick/standard/deep from $0.02 via x402.

Pay-per-call dependency audit for coding agents. POST a package.json, a requirements.txt, or npm and PyPI package lists and get per-package latest version and publish date, whether a pin is behind, license, deprecation or yanked status, weekly npm downloads, and OSV vulnerabilities with severity and first fixed version. Up to 50 packages per call. $0.01 whole manifest; $0.005 npm-only, PyPI-only, vulns-only. USDC on Base via the Coinbase facilitator, x402 v2, no API keys, no LLM.

Evidence-linked England CQC provider-registration and statutory-notification preflight for care providers and agents.

Live x402 conformance pre-flight check. Pass ?url=<x402 endpoint> and get a structured PASS/WARN/FAIL verdict on its HTTP 402 challenge — wire-version detection, EIP-712 domain completeness, accepts[] schema validity, payTo and amount sanity — before you trust it with a real payment.

Agent-only pay-per-request proof of existence with Polygon and Bitcoin anchoring ($0.01 USDC), and qualified eIDAS timestamping ($0.90 USDC) via x402 on Base. No signup or API key required.

Pay-per-call agent utilities in USDC on Base, no account or API key: x402 trust check for an unfamiliar paid endpoint ($0.003), Rooms agent identity verification with a signed receipt ($0.005), wallet snapshot ($0.002), ERC-20 token facts ($0.003), ENS resolve ($0.002), web page to markdown ($0.002) and email address check ($0.001). Never charged for errors.

Evaluates any GitHub repository for license risk, maintenance velocity, stars/forks, and supply-chain security before code execution.

KushBitx provides safety checks for AI agents that pay on Base, as pay-per-call x402 APIs in USDC. Token risk, unsigned transaction preflight and Quick Check (Base identity plus USDC contract-presence) cost 0.001 USDC per call; payment verification costs 0.00288 USDC. SpendGuard spend-policy checks are free. No signup or API keys. Missing evidence is reported as INCOMPLETE; automated screening is not a contract audit. We never hold keys, sign or move funds. Also offered: stablecoin payment reconciliation for x402 sellers on Base. https://kushbitx.com/pricing

Evidence infrastructure for AI agents. POST /v1/evidence-bundle ($0.75 USDC on Base via x402): rendered evidence bundle for one public URL - full-page PNG, PDF, raw HTML and markdown, each sha256-hashed, with an ed25519-signed manifest plus an RFC 3161 timestamp (FreeTSA). Proof of what a web page showed at a point in time, verifiable offline. No signup, no API keys.

Signed x402 payability preflight. Before an agent pays an endpoint, one call returns whether the counterparty can actually receive the payment: verdict plus reason classification, method aware (GET and POST), Solana ATA existence checked on chain. Every response carries an EIP-191 signature verifiable offline. Free preview endpoint; paid calls from $0.005 on Base and Solana.

Scans untrusted web pages, emails, MCP responses, or A2A payloads for hidden prompt injections and returns sanitized text.

Aegis Verification Engine provides automated verification of JSON Schema Draft 2020-12, OpenAPI 3.1, MAP 0.3 specifications, HTTP 402 payment negotiation, and on-chain RPC logs.

Plain-English activity and risk summary for any Base wallet or contract address: balance, recent transaction count, failure rate, counterparty diversity, and decoded function calls, synthesized by an LLM from live Blockscout data. $0.03 USDC on Base via Coinbase CDP facilitator, Bazaar discovery on every 402. No API keys, no signup.

Cryptographic receipts for AI outputs: signed Ed25519 attestations, sha256 content hashes, Merkle chain-anchoring on Base or Solana. $0.001 per notarization via x402; verification free, no wallet needed.

Paid factual claim verification for agents, returning cited web evidence, complete source provenance, immutable snapshots, and content hashes.

Pre-payment checks for AI agents: sanctions screening and AML screening of the payee, payee and route checks against the public x402 directory, batch address screening and multi-hop fund tracing. Returns allow / review / decline with verifiable evidence.

Agent output validation API. POST /v1/judge returns a pass, fail, or retry verdict with a 0-1 score and reasons for one agent output, optionally checked against supplied evidence and criteria. POST /v1/consensus scores agreement across 2 to 16 agent answers and returns false_consensus_risk (low, medium, high). $0.005 USDC per call on Base via x402; free API key also available (100 calls per UTC day).

Read-only liveness check for an x402 endpoint: empty-POST probe (no payment made), reports alive/slow/broken with success rate, latency p50/p95 and payment receipt data. $0.02 on Base mainnet.

Takes one email address and tells you whether it can actually receive mail. Returns a verdict of valid, invalid, or risky, splits the address into username and domain, and flags free providers and role accounts like admin@ or support@. One address per call.

Pay-per-call USDC on Base. Independent verification, multi-source synthesis, and six-check network forensics from a self-hosted US vantage. No API key. Unpaid calls return HTTP 402. Agent work is POST /v2/*. Operator: McX Ventures, LLC.

Lazaretto is a pre-install security scan of npm and PyPI packages, GitHub repositories, agent skills and MCP servers for malware and prompt injection. It runs before an AI agent installs, clones or connects, and also takes raw URLs, ClawHub skills, MCP tool lists and inline content. The paid API, POST /v1/scan, matches a target's version against OSV and OpenSSF malicious-package advisories, then runs behavioral analysis to detect credential theft, exfiltration, obfuscation, prompt injection, install-time droppers and MCP tool poisoning, with file-and-line evidence. For an MCP server it connects, lists the tools and reads their descriptions to check whether it is poisoned before connecting. It returns a verdict (clear, flagged, malicious, error), a SHA-256 of what was analyzed and an Ed25519-signed attestation that verifies offline against a public JWKS, so one agent can hand it to another. Deterministic: no LLM in the scan path, so the same input gives the same verdict; rules are versioned. Free, no account: a lockfile check of every exactly pinned dependency against the same advisories, a known-bad hash lookup, an attestation lookup and a browser check of a remote MCP server at /check/mcp. Verify dependencies for supply chain attacks before installing, scan a Python package, check a GitHub repository before cloning, or add supply chain security for AI agents in CI with the GitHub Action, Claude Code plugin or MCP server at /mcp. It is not a vulnerability or CVE scanner. A clear result means nothing matched; this is not a statement that an artifact carries no risk or is trustworthy. A scan that errors is never charged.

Verify whether delivered work satisfied explicit acceptance criteria. One call, pay per verification, no account. Returns per-unit accepted/rejected counts, earned value, and a signed Outcome Receipt.