Network · Hunt enumeration

Hunt enumeration: the first-100 funnels

Three enumeration crawls turn headline populations into huntable queues: the Registry live-filter, the Smithery-host re-verification and the ERC-8004 live-subset import. Each crawl is recorded as a dated snapshot whose per-stage exclusion funnel is itself a deliverable. The queue stays private; this page prints aggregates only.

The segments

• Segment B — Official Registry live-filter — status not_yet_crawled
  not yet crawled — enumerated count not measured yet
  Funnel: no crawl recorded yet
• Segment D — Smithery-host re-verification — status not_yet_crawled
  not yet crawled — enumerated count not measured yet
  Funnel: no crawl recorded yet
• Segment C — ERC-8004 live-subset import — status not_yet_crawled
  not yet crawled — enumerated count not measured yet
  Funnel: no crawl recorded yet

Counting rules

• Counting rule: enumeration uses public listing identifiers only — server names, registry namespaces and ERC-8004 slugs as printed in public listings; no contact data and no email addresses are collected or stored.
• Counting rule: the hunt queue is private — snapshot survivor identifiers are readable only by the account that recorded the snapshot; public surfaces print aggregates only, never a candidate identifier.
• Counting rule: a segment with no snapshot reads status not_yet_crawled with enumerated_n null — an uncrawled segment never prints an invented zero; zeros print as zeros only when a crawl actually ran.
• Counting rule (Segment C): servers ≠ accounts — listings_n counts surviving listings and operator_n counts distinct operator_ref values among them; the two numbers are never summed or substituted for each other.
• Counting rule (Segment C): the mass-mint exclusion removes every candidate whose domain accounts for strictly more than 10% of the snapshot's Segment C candidates (threshold 0.10); a domain at exactly 10% is not excluded, and the excluded domains and the excluded count print in the funnel.
• Counting rule: the public document composes, per segment, the latest snapshot by crawled_at across live accounts only (sandbox and canary accounts excluded), computed live at read time; reading it stores nothing.

The methodology

Hunt enumeration methodology: three enumeration crawls — the Official Registry live-filter (Segment B), the Smithery-host re-verification (Segment D) and the ERC-8004 live-subset import (Segment C) — are recorded as dated snapshots. Each snapshot applies its segment's deterministic filter to the candidate records the crawl observed (public listing identifiers plus probe and filter flags) and stores the per-stage exclusion funnel beside the survivor identifiers. The public document composes, per segment, the latest snapshot by crawled_at across live accounts only, computed live at read time; snapshot recording is the only write on this surface and public reads store nothing.

The limits

Limits, stated plainly: enumeration uses public listing identifiers only and stores no contact data and no email addresses. The hunt queue is private: survivor identifiers are readable only by the account that recorded their snapshot, and public surfaces print aggregates only — no candidate identifier ever appears in the public document, the page or the markdown twin. Snapshots are tenant-private records of what a crawl observed; this surface performs no crawl and no network fetch itself. A segment with no snapshot has not been crawled and says so, with a null count.

Callable surfaces

• POST /v1/hunt-enumeration/snapshots (account key required) — record one dated enumeration snapshot: segment B, D or C, a crawled_at ISO timestamp and at most 5000 candidates carrying public identifiers plus probe and filter flags; the stored record keeps the per-stage funnel and the survivor identifiers (the private queue)
• GET /v1/hunt-enumeration/snapshots (account key required) — the calling account's own snapshots only, each with its full funnel and survivor identifiers; another account's snapshots never appear
• GET /v1/hunt-enumeration (free, no key) — this public document: per segment, the latest snapshot across live accounts only, as aggregates (crawl status, crawled_at, funnel stages and enumerated_n; Segment C also prints listings_n and operator_n); a segment with no snapshot reads not_yet_crawled with enumerated_n null, and no candidate identifier appears anywhere in the document
• GET /hunt-enumeration (page) and GET /hunt-enumeration.md — the human-readable twin of the public document

Machine-readable document: GET /v1/hunt-enumeration · Validation gates: /validation-gates. Markdown twin: /hunt-enumeration.md.