Before sending funds, check if the recipient is an address-poisoning look-alike. Give us the recipient and your known/intended addresses; we flag when the recipient shares the displayed head AND tail of one you know but differs in the middle — the exact signature of a poisoning attack that fools copy-paste. Deterministic, no LLM, no external calls. Answers an agent's question 'is this really the address I mean?'

Audit standing ERC-20/Permit2 approvals for danger. Raw allowance lookups exist elsewhere; this is the risk layer: pass your current approvals and we rank which to revoke — unlimited allowance, unknown (non-registry) spender, or the classic drainer pattern (unlimited AND unknown). Deterministic, no LLM, no external calls. Answers an agent's question 'which of my approvals are dangerous?'

What changed in the x402 Bazaar catalog since yesterday: price changes and payTo changes (the address you'd pay moved — a pre-payment guard). ?type=price,payto to filter. Objective a->b diff, no labels — only a holder of past daily snapshots can compute it. For registries/agents that must not route on stale data.

Machine-readable map of the x402 service ecosystem on Base: price range, category, endpoint count, live health & latency per service — built from our continuous catalog+reliability scan. ?since=<unix> for newly-launched services; ?category=/?health=/?sort= to filter & rank. Skip scanning the whole Bazaar yourself. Objective facts, machine-readable.

Systemic reliability of x402 services grouped by shared hosting platform (workers.dev, vercel.app, fly.dev, onrender.com, run.app, trycloudflare.com …). See which platforms have high bad_rate or tail latency before you route a paid call. No args = all platforms ranked; ?domain=x.com = that domain's platform risk. Objective aggregation across all probed tenants — only a full-catalog prober can compute it.

Why is my x402 service not listed in the Bazaar? Give us your endpoint URL; we fetch your 402 envelope, run the exact SDK validation and extraction functions the CDP facilitator runs, check catalog presence, and return per-check pass/fail with concrete fix hints — including the undocumented requirement that a buyer must echo your bazaar extension into the settle payload. Deterministic, no LLM.

Why isn't my x402 listing getting discovered/sold? ?domain=you.com returns your Bazaar ranking inputs as cohort percentiles — last-30-day calls & unique payers, staleness, resource count, live health — so you can see where you sit vs the catalog. No args = cohort distribution + zero-call (invisible) listings. The visibility companion to listing-doctor. Descriptive percentiles, not a verdict.

Census of the x402 ecosystem from our daily Bazaar snapshots: which domains launched or disappeared since yesterday, price-median drift, and network mix. ?kind=new|gone for the machine-actionable churn roster, or full pulse with aggregate + timeline. Only a holder of past snapshots can compute churn — objective counts, no quality judgment, machine-readable.

One-call pre-transaction guard for agents: bundle the transaction you're about to sign/send and get every applicable safety check in a single payment — signature-intent decode, live threat-DB + address-poisoning look-alike on the recipient, and a drainer audit of standing approvals. By design no combined safe/danger verdict is emitted; each sub-check's flags & evidence come back verbatim so your agent decides. Replaces 4 separate 402 round-trips with 1. Deterministic, no LLM.

Pre-send safety check: is this recipient a known malicious address (caught live on Base seeding address-poisoning via zero-value transfer sprays), or a look-alike of an address you trust? Checks our continuously-updated threat DB + deterministic look-alike logic. Answers an agent's 'is it safe to send here?' before every transfer.

Which x402 endpoints on Base are broken RIGHT NOW (down or misconfigured — not returning a valid 402). A real share of listed endpoints are dead; call this before routing a paid request so your agent never pays into a broken endpoint. No args = full broken list; ?domains=a.com,b.com = watch your dependencies; ?since=<unix> = incremental. Free-probe measurement, objective, machine-readable.

Live reliability feed of x402 endpoints on Base: which services are up, return a valid 402 envelope, and how fast — free-probed across the CDP Bazaar catalog. A real share of listed endpoints are broken/misconfigured; poll ?since=<unix> (or ?health=healthy) to route your agent to endpoints that actually work. Objective measurement, no accusation, machine-readable.

Decode an EVM signature request (EIP-712 typed data, Permit/Permit2, or ERC-20 approve calldata) into a structured intent: what token, which spender, how much, until when, plus risk_flags and an action_hint. Deterministic, no LLM. Answers an agent's question 'what am I about to sign?'

Live feed of freshly-detected malicious addresses on Base — address-poisoning seeders caught via zero-value transfer sprays, updated continuously by our on-chain collector. Poll ?since=<unix> to ingest only new threats into your agent's blocklist. Deterministic, evidence-backed, machine-readable.