{"schema":"draconic21.signal_scorecard.v1","product":"Daily Agent Signal","started":"2026-10-05","method":"Hand-curated from archived Signal revisions. Each entry quotes a dated claim or falsifier we published before the outcome was known, then the later revision that tested it. Revisions are content-addressed: revision_sha256 is the sha256 of the exact brief bytes, so anyone holding that brief can check it. We log our own misses as falsified.","coverage_note":"Covers calls with an explicit, checkable test published from 2026-10-04 onward. Not every falsifier sentence in older editions is scored yet. Outcomes are about catalog and registration behavior (listings, re-registration timing, rolling 30-day quality counters), never settlements or seller revenue.","entries":[{"id":"relay402-10z-batch","claim":"relay402.georgespring.workers.dev re-registers its PayAI routes in short batches 8 hours apart, so the next batch lands near 10:00Z on 2026-10-05.","made_in":{"edition_date":"2026-10-04","generated_at":"2026-10-05T04:49:04Z","revision_sha256":"470f13060d260f711c90892496188264d831437e1df403c8dd2ba07e94abe0e3","signal_ids":["payai-head-relay402-8h-batch-rerank"]},"test":"A relay402 burst near 10:00Z; falsified by no burst or a burst at an unrelated time.","outcome":"confirmed","resolved_in":{"edition_date":"2026-10-05","generated_at":"2026-10-05T10:53:22Z","revision_sha256":"dda59220514b55c202530e203962b342d3eee418e6f0aa8dbecf0a8aa5acf052","signal_ids":["payai-head-relay402-8h-batch-rerank"]},"resolution":"11 routes re-touched at 10:00:59.701–10:01:08.691Z, the same 10:00Z route set as on 2026-10-03 and 2026-10-04."},{"id":"relay402-18z-shard","claim":"relay402 runs a fixed three-shard daily rotation; the 18:00Z shard (security-prompt-injection-risk, security-pypi-risk, security-secrets-check, six web-* checks, plus wayback-snapshot) re-touches near 18:01Z with the same route set.","made_in":{"edition_date":"2026-10-05","generated_at":"2026-10-05T10:53:22Z","revision_sha256":"dda59220514b55c202530e203962b342d3eee418e6f0aa8dbecf0a8aa5acf052","signal_ids":["payai-head-relay402-8h-batch-rerank"]},"test":"Falsified if the 18:00Z batch carries a different route set or misses the window.","test_after_utc":"2026-10-05T18:02:00Z","outcome":"open"},{"id":"secedgarterminal-45m-timer","claim":"The secedgarterminal.com route family re-touches on PayAI on a roughly 45-minute timer (second burst came 45m09s after the first).","made_in":{"edition_date":"2026-10-05","generated_at":"2026-10-05T05:54:45Z","revision_sha256":"12978b1ddbbf25f651efb53b8dfe47a03337713597996702fffbd1e3dba9ed29","signal_ids":["payai-head-secedgarterminal-solana-sec-burst"]},"test":"A third burst about 45 minutes after the second.","outcome":"falsified","resolved_in":{"edition_date":"2026-10-05","generated_at":"2026-10-05T07:58:42Z","revision_sha256":"b727c80922df02df147ed2f6b8a7e42039e44b14453822e801e8c3fbe82ab6a7","signal_ids":["payai-head-secedgarterminal-solana-sec-burst"]},"resolution":"No re-touch through 07:56Z, 2h43m after the 05:13Z burst. Our own timer reading was dropped."},{"id":"cdp-metadata-touch-reading","claim":"Fresh lastUpdated stamps with unchanged 30-day counters on CDP Base routes were metadata touches, not calls.","made_in":{"edition_date":"2026-10-05","generated_at":"2026-10-05T05:54:45Z","revision_sha256":"12978b1ddbbf25f651efb53b8dfe47a03337713597996702fffbd1e3dba9ed29","signal_ids":["social-trends-google-trends-thaw-2336z","financial-crypto-prices-held-24-9-post-thaw"]},"test":"Counters stepping later with lastCalledAt equal to the earlier lastUpdated would mean they were calls.","outcome":"falsified","resolved_in":{"edition_date":"2026-10-05","generated_at":"2026-10-05T07:58:42Z","revision_sha256":"b727c80922df02df147ed2f6b8a7e42039e44b14453822e801e8c3fbe82ab6a7","signal_ids":["cdp-quality-counter-lag-lastupdated-was-call"]},"resolution":"Six routes stepped by 07:56Z with lastCalledAt within 0.3s of the earlier lastUpdated. We published the correction and turned it into the counter-lag test below."},{"id":"cdp-counter-lag-five","claim":"Five pending lastUpdated stamps with flat counters are real calls that the catalog will count within about 8 hours, each with lastCalledAt equal to its stamp.","made_in":{"edition_date":"2026-10-05","generated_at":"2026-10-05T10:53:22Z","revision_sha256":"dda59220514b55c202530e203962b342d3eee418e6f0aa8dbecf0a8aa5acf052","signal_ids":["cdp-quality-counter-lag-lastupdated-was-call"]},"test":"Falsified for any stamp passing about 8h with no counter change, or a step whose lastCalledAt doesn't match a prior lastUpdated.","outcome":"partial","resolved_in":{"edition_date":"2026-10-05","generated_at":"2026-10-05T15:04:40Z","revision_sha256":"c72a97587ee6c849f2a4b2f0cc4aff5561132e4ffe760519d437335f2906ad2e","signal_ids":["cdp-quality-counter-lag-lastupdated-was-call"]},"resolution":"Four of five matched: x402financialdata crypto-prices 06:24:29Z, x402-stock-indicators crypto-prices 06:32:34Z, agent402 memory/remember 07:17:54Z, and losbeto research-brief 09:04:21Z against a 09:04:22Z stamp. google-trends /trend did not fit one-stamp-one-call: it jumped +43 with a later lastCalledAt (11:45:21Z)."},{"id":"trend-dip-decay-reading","claim":"google-trends /trend slipping 1,163/10 to 1,161/9 despite a new call was rolling-window decay; if /trend climbs back above 1,163 or payers return to 10 within a day, the dip was window noise instead.","made_in":{"edition_date":"2026-10-04","generated_at":"2026-10-05T00:50:02Z","revision_sha256":"e40f4026a7a1d35704814ed880db96b84332e5a1ddfa0962661b770b726eb5ea","signal_ids":["social-trends-google-trends-thaw-2336z"]},"test":"/trend above 1,163 calls or back to 10 payers within a day.","outcome":"falsified","resolved_in":{"edition_date":"2026-10-05","generated_at":"2026-10-05T15:04:40Z","revision_sha256":"c72a97587ee6c849f2a4b2f0cc4aff5561132e4ffe760519d437335f2906ad2e","signal_ids":["social-trends-google-trends-thaw-2336z"]},"resolution":"/trend reached 1,204 calls (lastCalledAt 11:45:21Z), so the decay reading failed its own test. Payers stayed at 9, not 10."},{"id":"payai-rallylive-testnet-flood","claim":"PayAI's +1,001 total jump (13,509 to 14,510) is mostly one seller's Base Sepolia testnet routes (rallylive.ca), not mainnet supply.","made_in":{"edition_date":"2026-10-05","generated_at":"2026-10-05T15:04:40Z","revision_sha256":"c72a97587ee6c849f2a4b2f0cc4aff5561132e4ffe760519d437335f2906ad2e","signal_ids":["payai-total-testnet-flood-rallylive"]},"test":"Falsified by a deeper read showing most of the +1,001 on Base or Solana mainnet from other hosts, or the rallylive rows switching to a mainnet network.","outcome":"open"}],"summary":{"tracked":7,"resolved":5,"confirmed":1,"partial":1,"falsified":3,"open":2},"how_to_verify":"Each made_in/resolved_in names an archived edition date, its generated_at, and the sha256 of that exact brief. If you hold that brief, sha256 it and compare. POST /v1/signal_day {\"date\":\"YYYY-MM-DD\"} ($0.04) returns the final stored edition for a date, which can be a later same-day revision than the one cited, so hash-matching needs the brief you bought at the time. Today's brief is POST /v1/signal_latest ($0.03).","honesty":"Scored claims concern catalog and registration behavior, not settlements or seller revenue. Listings≠revenue.","links":{"health":"https://draconic21-x402-api.onrender.com/health","example":"https://draconic21-x402-api.onrender.com/v1/signal_latest/example","buyer":"https://draconic21-x402-api.onrender.com/buyer.md"}}