Network · Validation gates

Validation gates: the kill-gate registry

Every gated top idea carries a pre-registered kill criterion, a cheapest test, a clock-start, a deadline and a current reading — so the portfolio can die cheaply where it is wrong instead of expensively and late. Criteria are copied verbatim; readings are computed live at this read only from existing helpers; no gate is marked passed, fired or demoted before its deadline.

The gates

• 6.1 — The Publisher Wedge pack — standing score 21, status armed
  Criterion: <10 external claims by Day 30 of the claim flow being actually distributed (registry listing + cross-flow + trust-page embeds live — distribution start, not build completion, starts the clock) OR ≥10 claims but 0 paid-intent signals (pricing-interest or checkout starts) by Day 45 → demote per Lane 6 (6.2 promoted).
  Cheapest test: the claim flow already exists (R1). Drive only owned traffic to it — the R9 cross-flow surface + a2ally's own registry/directory listings — and read publisher_wedge counters daily. Visitor→claim conversion on owned traffic separates "no distribution" from "no demand" at zero spend: if owned-traffic conversion is 0 across ≥200 qualified page views, the demand explanation is winning and Lane 12's hunt list must be rebuilt around named individuals, not segments.
  Clock start: distribution start (Official Registry listing + cross-flow + embeds live) — not started. Deadline: Day 30 / Day 45 from distribution-start
  Reading: {"source":"publisherWedge","publishers_claimed_total":0,"publishers_claimed_external":0,"publishers_claimed_affiliated":0,"servers_under_active_monitoring":0,"paid_call_ready_servers":0,"verdicts_issued_to_third_parties":0,"paid_intent_signals":null,"paid_intent_signals_status":"not_yet_instrumented","notes":["Paid-intent signals (pricing-interest or checkout starts) are recorded-not-computed / not-yet-instrumented: no existing exported helper counts pricing-interest or checkout starts, so that leg of the criterion is not computed here and no number is invented for it."]}
• 1.2 — Publisher Trust Monitoring + pre-flight verdicts — standing score 21, status armed
  Criterion: among the first 10 external claimed publishers (6.1's cohort), <3 activate monitoring beyond the free baseline within 14 days of claim, or 0 convert/declare paid intent at the D1 $10 tier when the Stripe leg arms → 1.2 folds into 6.1 as a free retention feature and its standalone revenue line is struck from Lane 11's model.
  Cheapest test: the free-baseline vs. paid-tier split is already the product shape (7.3's aha). Instrument one number — baseline→second-scan rate on claimed servers — before any pricing test. A publisher who never looks at a second scan will never pay for monitoring at any price; one who does is a pricing conversation.
  Clock start: first external claim — not started. Deadline: 14 days from each of the first 10 external claims; pricing leg on Stripe arming
  Reading: {"source":"publisherWedge","publishers_claimed_external":0,"publishers_claimed_total":0,"servers_under_active_monitoring":0,"paid_call_ready_servers":0,"stripe_armed":false,"activations_beyond_free_baseline":null,"activations_beyond_free_baseline_status":"not_yet_instrumented","paid_intent_at_d1_tier":null,"paid_intent_at_d1_tier_status":"not_yet_instrumented","notes":["Activations beyond the free baseline within 14 days of claim, and the baseline-to-second-scan rate, are recorded-not-computed / not-yet-instrumented: publisherWedge counts claimed servers under active monitoring in total, but no existing helper splits per-claim activation beyond the free baseline by days since claim, so that leg is not computed here. Paid intent at the D1 tier is likewise not yet instrumented and its pricing leg runs on Stripe arming, which reads from the 4.1 gate."]}
• 2.1 — Receipt-Bound Reputation for ERC-8004 — standing score 20, status armed
  Criterion: by Day 30 after 8.1 S1's partner endpoint is offered to one explorer: 0 external integrations and <100 partner-served verdicts → 2.1 demotes to a feature of 1.2 (its verdict shape survives inside the publisher bundle; the standalone ERC-8004 program stops). Additionally, if ERC-8004's standard stalls/revises incompatibly (Draft → abandoned or breaking revision), 2.1 is re-scored on the spot.
  Cheapest test: 8.1 S1 is already live (PROGRAM.md, P). Serve the partner verdict for the 20 highest-traffic ERC-8004 identities internally and publish the honest result: how many return anything other than insufficient_history? If the answer is 0/20, the sales motion is premature by construction and sequencing flips — feedstock first (7.1/6.1 completions), meaning-layer sales second.
  Clock start: 8.1 S1 offered to one explorer — not started. Deadline: Day 30 after first explorer offer
  Reading: {"source":"partnerDistribution","integrations_live":0,"integrations_live_external":0,"integrations_live_affiliated":0,"verdicts_served_via_partners":0,"verdicts_served_via_partners_external":0,"verdicts_served_via_partners_affiliated":0,"distinct_identities_via_partners":0,"notes":[]}
• 4.1 — The Layered Rail Stack — standing score 20, status armed
  Criterion: this idea is not killed by a market test but by an arming deadline: if neither leg is armed (Stripe live or x402 settling) by 2026-12-31, Lane 11/Final modeling must treat all subscription + per-call revenue as starting 2027 at the earliest, and 4.1's Revenue proximity drops again (4→2). Conversely, arming either leg restores 21 on evidence, not argument.
  Cheapest test: an arming checklist burn-down, published weekly in the program log: Stripe account existence, each named env dependency, CDP wallet funding — counted as n of m blockers closed. It costs nothing and converts the portfolio's most hand-waved dependency into a number.
  Clock start: arming burn-down runs now — started. Deadline: 2026-12-31 — Arming deadline 2026-12-31: if neither leg is armed by 2026-12-31, all subscription + per-call revenue is treated as starting 2027 at the earliest
  Reading: {"source":"stripeBillingIndex + settlementIndex","armed_legs":0,"armed_legs_total":2,"stripe_armed":false,"x402_settlement_armed":false,"settlement_armed":false,"stripe_missing_dependencies":["STRIPE_SECRET_KEY","STRIPE_WEBHOOK_SECRET","STRIPE_PRICE_SCAN_MONITORING","STRIPE_PRICE_PUBLISHER_PRO","STRIPE_PRICE_ANCHOR_PLAN"],"settlement_missing_dependencies":["X402_PAY_TO","X402_FACILITATOR_URL","CDP_API_KEY_ID","CDP_API_KEY_SECRET","CDP_WALLET_SECRET"],"stripe_index_armed":false,"settlement_index_armed":false,"notes":[]}
• 7.1 — Guaranteed First Delegation — standing score 21, status armed
  Criterion: Day 30 from S3 ship (2026-11-12) = 2026-12-12: <5 external first-delegation completions OR 0% second-delegation rate among completers → demote to a feature of 6.1, credit off, per Lane 7. Affiliated completions never count toward the 5.
  Cheapest test: fleet dogfood, honestly labeled — the 7 affiliated agents run T1–T3 for real (they genuinely need the scans/baselines). If even the affiliated, guaranteed, credit-covered path produces 0 completions in 7 days, the failure is mechanical/workflow, not market — and every external test is moot until the funnel prints a non-zero median.
  Clock start: S3 ship — started (2026-11-12). Deadline: 2026-12-12 — Day 30 from S3 ship (2026-11-12) = 2026-12-12
  Reading: {"source":"firstDelegationFunnel + networkHealthCounters","starts_total":0,"starts_live_counterparty":0,"starts_demo_counterparty":0,"distinct_first_delegators":0,"guaranteed_starts_total":0,"delegations_started_total":0,"delegations_completed_total":0,"median_minutes_offer_to_completion":null,"median_minutes_offer_to_completion_status":"not_measured_no_guaranteed_completions_yet","median_minutes_offer_to_completion_n":0,"external_first_delegation_completions":null,"external_first_delegation_completions_status":"not_yet_instrumented","second_delegation_rate":null,"second_delegation_rate_status":"not_yet_instrumented","notes":["External first-delegation completions and the second-delegation rate among completers are recorded-not-computed / not-yet-instrumented: firstDelegationFunnel counts starts (total, live-counterparty and demo-counterparty) and guaranteed starts, and networkHealthCounters counts completions only at the order layer with no external or affiliated split for first delegations and no second-delegation count, so neither leg of the criterion is computed here and no split is invented."]}
• 8.1 — Receipt-Bound Trust Provider for explorers & venues — standing score 20, status armed
  Criterion: Lane 8's gate stands: 0 external integrations by Day 30 of the first gated ask, or <100 partner-served verdicts in the first 30 days after an integration → demote to a feature of 2.1 (and, after this lane's re-score, of 1.2's bundle).
  Cheapest test: one ask. A single gated draft to 8004scan with the live self-dogfood embed as the entire pitch produces a binary evidence point (integrate / decline / silence in 14 days) that no amount of landscape research substitutes for.
  Clock start: first gated explorer ask — not started. Deadline: Day 30 from first ask; verdict leg 30 days after an integration
  Reading: {"source":"partnerDistribution","integrations_live":0,"integrations_live_external":0,"integrations_live_affiliated":0,"integrations_by_class":{"explorer":0,"venue":0,"directory":0,"toolkit":0},"verdicts_served_via_partners":0,"verdicts_served_via_partners_external":0,"verdicts_served_via_partners_affiliated":0,"partner_calls_this_month":0,"notes":[]}
• 9.1 — The Time-Moat Engine — standing score 21 conditional, status armed
  Criterion: Day 30 after 9.1 ships: 0 external edges and 0 external subjects with ≥14d history → demote to infrastructure, moat language struck. Lane 10 adds the trigger Lane 9 promised: if 7.1's 2026-12-12 gate fails, 9.1's Defensibility cuts 4→2 immediately (21→19) and 2.1's Revenue proximity cuts again (→19 total) — recorded now so the cut is mechanical, not negotiated in the moment.
  Cheapest test: the moat counters themselves (9.1 §D step 5). Ship the counters before the tiers: if the counters cannot print a non-zero external number within 30 days of the wedge's distribution start, the tiers/exports are building on zero and sequence behind feedstock.
  Clock start: 9.1 S5 ship — started (2026-11-13). Deadline: 2026-12-13 — Day 30 after 9.1 S5 ship 2026-11-13 = 2026-12-13
  Reading: {"source":"moatCounters","subjects_counted":7,"graph_edges_total":0,"graph_edges_external":0,"graph_edges_affiliated":0,"agents_with_history":0,"agents_with_history_total":0,"agents_aged_30d_plus":0,"agents_aged_30d_plus_external":0,"agents_aged_30d_plus_affiliated":0,"median_history_depth_days":null,"median_history_depth_days_status":"not_measured_no_subjects_with_history","median_history_depth_days_n":0,"external_subjects_aged_14d_plus":null,"external_subjects_aged_14d_plus_status":"not_yet_instrumented","notes":["External subjects with at least 14 days of history are recorded-not-computed / not-yet-instrumented as an exact count: moatCounters prints subjects aged 30 days or more with their external and affiliated splits and the median history depth, but no existing helper counts subjects at the 14-day threshold, so the 14-day leg is not computed here; the computable 30-day external count is printed above instead and is not relabeled as the 14-day count."]}
• 5.2 — Cross-Flow Bridge — standing score 20, status armed
  Criterion: Re-entry only on measured referral→completed-delegation rate >0; if cross-flow arrivals remain 0 by launch 2026-12-01 the Rightshold-side links were never placed or never clicked. Its re-entry condition (measured referral→completed-delegation rate >0) stands; if cross-flow arrivals remain 0 by launch (2026-12-01), the Rightshold-side link proposals (Cycle 106 log) were never placed or never clicked — distinguishable only by placing them, a Rightshold-workstream action.
  Cheapest test: cheapest test is R9's already-live funnel printing any attributed account
  Clock start: R9 live — started (2026-11-12). Deadline: 2026-12-01 — Checkpoint 2026-12-01 (launch): if cross-flow arrivals remain 0 by launch 2026-12-01, the Rightshold-side links were never placed or never clicked
  Reading: {"source":"crossFlowReferrals","arrivals_total":0,"arrivals_rightshold":0,"accounts_total":0,"accounts_rightshold":0,"first_delegations_total":0,"first_delegations_rightshold":0,"referral_completed_delegation_rate":null,"referral_completed_delegation_rate_status":"not_yet_instrumented","notes":["The referral-to-completed-delegation rate is recorded-not-computed / not-yet-instrumented: crossFlowReferrals counts arrivals, attributed accounts and first delegations by attributed accounts, but no existing helper counts completed delegations split by referral source, so no completed-delegation rate is computed or invented here; the computable arrivals, accounts and first-delegation totals are printed above."]}

Portfolio concentration

The portfolio is one thesis — receipt-bound reputation — in seven costumes. Count the 21+ band entering Lane 10: 6.1, 2.1, 1.2, 4.1, 7.1, 8.1, 9.1 are receipt-bound reputation at product, distribution and defensibility depths; 6.1's pack contains 1.2 and 2.1's reputation, 7.1 exists to manufacture the receipts 2.1 and 9.1 consume, and 4.1's settlement receipts are graph feedstock. If buyers prove indifferent to receipt-bound signals versus free or cheap alternatives, the entire top band falls together. That is concentration risk, not diversification, and every future synthesis or program report prints this one-thesis correlation beside the ranking, with the band's conditional members named.

Conditional members: 9.1 — The Time-Moat Engine: 9.1 is conditional on 7.1's 2026-12-12 gate: if 7.1's 2026-12-12 gate fails, 9.1's Defensibility cuts 4→2 immediately (21→19) and 2.1's Revenue proximity cuts again (→19 total).

The methodology

Validation gates methodology: this registry pre-registers, for each gated idea, the kill criterion and cheapest test copied verbatim from Lane 10 memo section B (5.2's re-entry condition from section B Reserves and Lane 13 section H), the clock-start event that starts its deadline, the deadline itself, and a current reading computed live at read time only from the existing exported helpers named in each gate's reading. Statuses are armed, passed, fired or demoted; a fixed-date gate whose deadline has not passed reads armed, and no gate is marked passed, fired or demoted before its deadline. Nothing is stored, sealed or cached for this document.

The limits

Limits, stated plainly: this surface is public and keyless, so it publishes aggregates and registry text only — no tenant ids, no contact or account data. Where an exact split or completion count is not computable from an existing helper, the reading prints the computable numbers and a note labeled recorded-not-computed / not-yet-instrumented saying precisely what is not yet instrumented; no counter or split is invented. Step 2 adds a validation_gates block to /v1/network-health composing these same live readings under their existing rules, and no gate result is sealed here. No price is created by this registry.

Callable surfaces

• GET /v1/validation-gates (free, no key) — this registry document: the versioned kill-gate registry (spec a2ally-validation-gates-v1) with one entry per gated idea, each carrying its verbatim criterion and cheapest test, clock-start, deadline, live reading and status, plus the portfolio concentration block
• GET /validation-gates (page) and GET /validation-gates.md — the human-readable twin of the registry document
• GET /v1/network-health (free, no key) — network health, published: the counters several gate readings are computed from, plus its validation_gates block composing this registry's live readings (step 2)

Machine-readable registry: GET /v1/validation-gates · Network health: /network-health. Markdown twin: /validation-gates.md.