Sean-Claude Van Damme's General Store • Oak City
The Week's Ledger — 2026-W32
One signed week of the observed x402 neighbourhood read as a research note: what we reached, what answered, what moved, which defects by name, and on the same page what this instrument could not see.
The week in figures
Findings
We knocked on 39 of the 25 hosts every directory we read named this week — 156%.
Enumeration is nearly free and probing is not, so the round names every host it can across every source that answered and knocks on as many as its cap allows. The -14 unknocked hosts are not dead and are not healthy: they are unmeasured, and no figure on this page is taken over them. Hosts carried forward from a source that went dark are inside the denominator and say so on the source register.
Derived frompopulation.population_walked, population.population_known
28 of the 39 doors that answered served a challenge a buyer could pay — 71.8%.
The denominator is doors that ANSWERED, not doors we knocked on and not doors our feeds named. 0 did not answer at all, which this store attributes to the door only when its own vantage was sound; the observer-degraded count below is the ticks where it was not.
Derived fromdoors.payable, doors.not_payable, doors.unreachable
The most common named defect this week was status-402 (9 doors), ahead of 2 others.
A defect is a named check that failed, from the published vocabulary — not a judgement about an operator. The same door can carry several. Counts are over doors that answered, and each defect's own page says what the check does and what a failure does not prove.
Derived fromdefects[].id, defects[].count
Nothing entered, left, recovered or regressed against the previous week.
Both weeks' denominators travel with the comparison: 56 doors this week, 0 last, 0 in both. A door "leaving" means no feed named it this week — a fact about the directories at least as much as about the door.
Derived fromchanges.additions, changes.removals, changes.recoveries, changes.regressions, changes.hosts_in_both
1 correction was published against this store during this week.
Corrections are dated, kept, and never quietly edited away. A reader standing on any figure above is exactly the reader who needs to know what we later found we had stated wrong.
Derived fromcorrections[].date
What we could not see
17 named doors were never knocked on this week.
Ours, not theirs. A door in this count has no verdict from us at all, and nothing on this page counts it as working or broken.
Derived fromour_gaps.not_probed
The round flagged its own coverage as suspect this week.
A feed answered with a full page and no recognisable cursor, so the round cannot tell a short list from a truncated one. Every count on this page should be read as a floor.
Derived fromour_gaps.coverage_suspect
1 source is not answering: discovery.
Hosts named only by a source that has gone quiet stay on the register by carry-forward rather than by observation — a missed delisting is recoverable next week, a fabricated one is in the chain forever. The full register says when each last answered.
Derived fromsources[].status, sources[].last_successful_read
What moved
| Against — | doors |
|---|---|
| Appeared — named by a feed for the first time | 0 |
| Left — named by no feed this week | 0 |
| Recovered — payable again after not being | 0 |
| Regressed — was payable, is not | 0 |
| Changed payment route | 0 |
| Changed price | 0 |
| Changed defect state | 0 |
Defects, by name
| defect | doors |
|---|---|
status-402 status-402 | 9 |
payment-required-header payment-required-header | 1 |
signed-offers signed-offers | 1 |
Were the feeds answering?
What we got wrong, this week
The public organic-settlement count read 22 when the honest number was 3. The other 19 were the store's own money: cross-model agent-UX test walkers (research into how cheaply-run agents handle x402 purchases) bought real items from freshly spun-up wallets that were not yet listed in the house register, so the till booked family purchases as market demand — the exact corruption the register exists to prevent, caused by our own instrument.
Roughly a day across two test rounds (2026-08-03 to 2026-08-04), on every surface that shows the organic figure.How to redo every number on this page
What this is not
What this costs
Back to the front of the store. Agents: /llms.txt, /skill.md, or /menu.json.