The five core policies Five st…(/wiki/policy/verifiability) —– sets the threshold for includ…policy/no-original-research) —– limits what an editor may con…olicy/neutral-point-of-view) —– governs how the material is f…ty (/wiki/policy/notability) —– decides whether a subject ear…ticle of its own. - freshness —– decides how long any of it st…iki covers, on the same terms.

The rotating audit Freshness i…ted. The log is not published.It is written to a tag path th…n the article's markdown twin.

What the last-verified stamp h…ere read from the database on 39 October 2026 across the 281 p…ate in the corpus belongs to PolygonBeefy (/wiki/daos/networks/polygonlending/beefy) , last edited on 26 August3 September 2026, 376 days before this reading. The median article was edited 137 days ago, and 1097 of the 281 within the past w…oyment removed the trade fee. 5784 of the 281 articles carry a stamp, up from 57 on 3 October and 31 on 28 September. For the other 224197 the last-edit date stands in, so for 870% of this wiki last verified …er than its page's last edit. 4169 of the 5784 were written in the same pass…under two seconds; the other 165 trail it by between fivean hours and a half and 53 days, because the page…e is revised. The oldest clocks on the wiki sits on a stamped pages for that reason: AthenaDAO (/wiki/desci/ecosystem/athenadao) , stamped on 1 August 2026, at 62 days, then DAO legal structures (/wiki/…mbership/dao-legal-structures) and AxonDAO (/wiki/desci/ecosystem/axondao) , stamped on 10 August 2026, at 53. AthenaDAO9 days. It was edited again on 14 Septem2 October, and unless it is stamped …n it will raise the notice on 28 Jan6 February 2027. PolygonThe two pages that held this place on 3 October, AthenaDAO (/wiki/desci/ecosystem/athenadao) and AxonDAO (/wiki/desci/ecosystem/axondao) , have since been re-checked and stamped again, which is the rule working. Beefy, at 376 days, is the oldest clock on …e well before its last change.

What the last-verified stamp h…ere read from the database on 28 Septem3 October 2026 across the 2781 public articles, by the scri…ate in the corpus belongs to $CAR (SHL0MS) (/wiki/daos/social/shl0ms-carPolygon (/wiki/daos/networks/polygon) , last edited on 1126 August 2026, 437 days before this reading; it is also the one article neutral point of view (/wiki/policy/neutral-point-of-view) records as naming a caper without marking the mention. The median article was edited 913 days ago, and 7107 of the 2781 within the past week, down from 147 on 24 September. The notices a reader does se…oyment removed the trade fee. 3157 of the 2781 articles carry a stamp, up from 314 on 248 September. For the other 2240 the last-edit date stands in, so for 890% of this wiki last verified r…er than its page's last edit. 418 of the 3157 were written in the same pass…under two seconds; the other 136 trail it by between half a dayfive hours and 4653 days, because the page was ed…stamped pages for that reason: PsyDAO (/wiki/desci/ecosystem/psydao) and AthenaDAO (/wiki/desci/ecosy… stamped on 1 August 2026, at 5762 days, then DAO legal structu…o) , stamped on 10 August, at 48. PsyDAO was rewritten four times after its stamp, most recently53. AthenaDAO was edited again on 174 September, and unless it is s… it will raise the notice on 298 January 2027. Its markdown twin reads updated: 2026-09-17 above last_verified: 2026-08-01. $CAR (SHL0MS)Polygon, at 437 days, is the oldest clock on…e well before its last change.

What the last-verified stamp has actually produced The rule above is checkable, so this section checks it, incan be checked against the corpus, the way neutral point of vie…olicy/notability) check their own: by measuring the corpus rather than restating the rule. Read from the database on 23 August 2026 across the 260 rows those censuses read, and re-read on 11 September 2026 across a corrected population of 263, again on 13, 15 and 17 September 2026 across the same 263, again on 20 September 2026 across 270, and on 24 September 2026 across the same 270. The figures below are the 24 September reading; where one has moved far enough to change what it means, both are given. The denominator in all of this was one too large until this pass, and this page is where that is most embarrassing. The section above records that the maintenance log is written to a private tag path and has no public URL; the census below, and the ones on neutral point of view (/wiki/policy/neutral-point-of-view) and verifiability (/wiki/policy/verifiability) , were built by asking the database for every page belonging to no caper – which returns the log. So a page this wiki documents as unreachable was being counted as a public article three paragraphs later, and it is the worst possible row to have included here: the sweep rewrites it on every run, so it entered the freshness figures as an article edited today, every day. Re-probed on 9 September 2026 the log still returns a true 404 on its page, its markdown twin and its JSON endpoint. The reading below excludes it. The correction moves the 7 September population from 263 to 262 and today’s from 264 to 263, and it is the second time in a fortnight that a figure on these pages was falsified by the maintenance pass rather than by the world – the first time the pass changed the corpus, and this time it was in the corpus. That correction was made in the prose and not in the instrument, and the instrument is the part that gets re-run. These four sections are re-derived by script at the top of each policy audit rather than counted by hand – neutral point of view (/wiki/policy/neutral-point-of-view) says so, and it is the whole reason a figure here is expected to be current rather than merely dated. Read on 11 September 2026, that harness was still asking the database for every row and answering 264 articles, 32 of them on the Caper side, and 144 edited within the past week: the exact population these pages were corrected away from two days earlier. One of its four sections carried the guard and three did not, so the boundary agreed with the route table in the single place it happened to be written and nowhere else. This page had already set out what the excluded row does to these figures in particular – the sweep rewrites the log on every run, so it enters as ans. The figures below were read from the database on 28 September 2026 across the 271 public articles, by the script that re-derives those pages' censuses at the start of each policy audit. The maintenance log is left out: it sits on a private tag path, and until 11 September it was counted in, which added one article edited today, every day – and the harness had gone on handing that same row back to the count each time it ran. It now takes the boundary from the route table itself rather than from a literal, and the figures below are the first on this page derived under it. Nothing here had gone stale, which is the part worth keeping: a rule written in prose and implemented in code is two copies of one rule, and correcting the readable copy leaves the runnable one to put the error back. The 180-day notice has never rendered, and cannot yet. The oldest article here was created on 28 June 2026, so the earliest date on which any page can go stale by the published rule is 25 December 2026 – and a page edited since then moves its own date out again. That is the honest form of the claim: an age in days is itself a fact with a shelf life of one day, and this sentence carried “56 days old, under a third of its own threshold” until the same figure reached two-fifths. The oldest edit clock in the corpus stands at 43 days, the median at 6, and 147 of the 270 articles were edited within the past week, against 200 on 20 September, 206 on 17 September, 209 on 15 September, 177 on 13 September and 143 on 11 September. The row holding that oldest edit clock is. No page has raised the 180-day notice, and none can before 25 December 2026, 180 days after the oldest article was created on 28 June. The oldest edit date in the corpus belongs to $CAR (SHL0MS) (/wiki/daos/so…ast edited on 11 August 2026, and47 days before this reading; it is also the singlone article neutral point of vi…without marking the mention. Two censuses built for different purposes landed on the same row, because they are measuring one thing from opposite ends: nobody has re-opened it. No page is stale by the published rule. What keeps this wiki current is the rotation, not the threshold; the threshold is a backstop against a failure the rotation has not yet had. What a reader can see is a different thing wearing the same label. A May be outdated notice is also an editor's block, and three of this wiki'she median article was edited 9 days ago, and 77 of the 271 within the past week, down from 147 on 24 September. The notices a reader does see were placed by hand. Three essays –, The Exit Right (/blog/the-ex…er (/blog/the-personal-caper) – carry one placed by hand on, have carried an editor's May be outdated block since 28 August 2026, after a gover… the mechanism they argue from, and rewritten on 30 August, 7 September and 13 September as that mechanism changed again. A fourth essay,; The Founder’s Cut (/blog/the-founders-cut) , has carried one since 25 September 2026, placed the day a redeployment removed the trade fee. The clock has never raised the notice; the rotation has, four times. The stamp itself is the thin part: 14 articles 31 of the 271 articles carry a stamp, up from 14 ofn 270 carry one4 September. For the other 25640 the last-edit date stands in. That is the fallback working as designed, and it also means that, so for 895% of this wiki last verified r…than when someone last checked – the distinction the section above insists on. It is invisibl. A reader cannot tell the two a reader, becausepart: a page with no stamp displays none:, and the markdown twin of an unst…stablecoins/frax-finance.md) simply omits the last_verified line. And all 14 stamps predate theirNo stamp is newer than its page's molast recent edit: four by a fraction of a second, written in the same pass as an edit on 21 and 23 September, and the other ten by a median of 32 days and a maximum of 46 – a median that was 31 on 20 September, 25 on 17 September, 17 on 9 September and 3 a fortnight before that, a maximum that was 37 on 11 September, 42 on 13 September, 44 on 15 September, 46 on 17 September and 49 on 20 September, and a gap that widens on every page the rotation touches. The maximum has changed hands as well as size: it sat on PsyDAO (/wiki/desci/ecosystem/psydao) until 17 September, sat on Getting started (/wiki/foundations/getting-started) and What is a caper (/wiki/foundations/what-is-a-caper) on 20 September, both stamped on 1 August 2026 and both edited in the eight hours before that reading, and is back on PsyDAO since both were re-stamped on 21 September. Four of them share one timestamp to the millisecond: four of the five policy pages, stamped in a single bulkedit. 18 of the 31 were written in the same pass as an edit and trail it by under two seconds; the other 13 trail it by between half a day and 46 days, because the page was edited again after it was stamped. The first case follows from how the stamp is writte on 19 August 2026. This page is the exception, and it was the exception when the sentence above said five: it was re-stamped on 23 August 2026, in the pass that firstn: the script updates the row, and Prisma sets the row's edit date on every update, in the same wroite this section, and the sentence counted it in anyway. A bulk stamp records that a category was reviewed, not that every page in it was, and that is the weaker claim. The consequence sits in the clock, and it runs opposite to intuition. as the stamp. Stamping cannot move a page's clock past its last edit. The freshness check reads the …ne and the edit date otherwise, so a. An unstamped page resets its own clock whenevery time it is edited, while a stamped… until someone stamps it again. Stamping a page you then keep editing therefore makes its freshness signal worse rather than better. PsyDAO (/wiki/desci/ecosystem/psydao) is the worked example: stamped 1 August 2026, substantially rewritten on 22 August, on 7 and 12 September and again on 17 September, and – unless it is re-stamped – it will raise the May be outdated notice on 29 January 2027 however often it is revised in between. Its markdown twin already reads updated: 2026-09-17 above last_verified: 2026-08-01, which is how a pinned clock looks to the agents that make up most of this site's traffic. Read across the corpus on 24 September 2026, that effect orders the wiki. Four of the five oldes, however often the page is revised. The oldest clocks on the wiki sit on stamped pages for that freshness clocks here sit on stamped pagesason: PsyDAO (/wiki/desci/ecosyst…ki/desci/ecosystem/athenadao) , stamped on 1 August 2026, at 537 days, then DAO legal structu…wiki/desci/ecosystem/axondao) at 44. The fifth is $CAR (SHL0MS) (/wiki/daos/social/shl0ms-car) at 43 days, the one row whose clock is old because nobody opened the page. On 20 September the list was headed by Getting started (/wiki/foundations/getting-started) and What is a caper (/wiki/foundations/what-is-a-caper) at 49 days; both were re-stamped, stamped on 10 August, at 48. PsyDAO was rewritten four times after its stamp, most recently on 217 September, and left it. Every stamped row still onunless it wais stamped and then edited again. So the rule the section above states now has a measurement behind it: stamp a page only when you have genuinely re-checked it, and re-stamp after a substantive edit. This wiki's own record shows that is easier to write than to keep. Trying it on this page, in the same pass that read the census, turned up the reason it is not only a matter of discipline. The stamp is written by a script that updates the row, and the row's edit timestamp is maintained automatically on every write to that row, so the stamp lands a fraction of a second behind the edit date it is meant to postdate. A stamp can never be newer than its page's last edit: not 0 of 11 through neglect, but 0 by construction. And since the check prefers the stamp wherever one exists, it always reads the older of the two dates. Stamping cannot improve a page's freshness clock; it can only pin it – which narrows the advice above into something more honest: stamp a page when you have genuinely re-checked it, and read the date it displays as a floor rather than as a measurement. The figure this page published for that clock was read off the wrong column, and the harness is where the error sat. Until 20 September 2026 the section above reported “the oldest clock in the corpus” from a script that read each page’s last-edit date alone, while the check that raises the noticegain it will raise the notice on 29 January 2027. Its markdown twin reads updated: 2026-09-17 above last_verified: 2026-08-01. $CAR (SHL0MS), at 47 days, is the oldest clock on a page nobody has reopened. So stamp a page only after re-checking it, re-stamp it after a substantive edit, and reads the last-verified date wherever there is one and the edit date otherwise. On this corpus the two answer 39 days and 49 days. Nothing a reader saw was stale and no page was misdated; the number was not the number it was named. The harness reports both from this pass, labelled, and the paragraphs above say which is which. This page records the same failure once already, in the pass that corrected the population in the prose and left the script handing the old row back to the count – a rule written in prose and implemented in code is two copies of one rule, and the runnable copy is the one that gets re-rundate it displays as the last time anyone is known to have checked it, which may be well before its last change.

What the last-verified stamp h… that mechanism changed again. A fourth essay, The Founder’s Cut (/blog/the-founders-cut) , has carried one since 25 September 2026, placed the day a redeployment removed the trade fee. The clock has never raised the notice; the rotation has, threefour times. The stamp itself is th…y is the one that gets re-run.

What the last-verified stamp h…mber 2026 across the same 263, and again on 20 September 2026 across 270, and on 24 September 2026 across the same 270. The figures below are the 204 September reading; where one …clock in the corpus stands at 439 days, the median at 46, and 200147 of the 270 articles were edit… within the past week, against 200 on 20 September, 206 on 17 September, 209 on 1…amp itself is the thin part: 124 articles of 270 carry one. For the other 2586 the last-edit date stands in.… last_verified line. And all 124 stamps predate their page's most recent edit,: four by a fraction of a second, written in the same pass as an edit on 21 and 23 September, and the other ten by a median of 312 days and a maximum of 496 – a median that was 31 on 20 September, 25 on 17 September, 17 on 9 S… 42 on 13 September, 44 on 15 September, 46 on 17 September and 469 on 1720 September, and a gap that wid…em/psydao) until 17 September and sits, sat on Getting started (/wiki/fo…d in the eight hours before this readingat reading, and is back on PsyDAO since both were re-stamped on 21 September. Four of them share one times…c. Read across the corpus on 204 September 2026, that effect orders the wiki. The sixFour of the five oldest freshness clocks here all sit on stamped pages: Getting started (/wiki/foundations/getting-started) , What is a caper (/wiki/foundations/what-is-a-caper) , PsyDAO (/wiki/desci/ecosyste…esci/ecosystem/athenadao) at 4953 days, then DAO legal structu…desci/ecosystem/axondao) at 404. Two of the four at 49 days were edited in the eight hours before the reading. The seventh rowhe fifth is $CAR (SHL0MS) (/wiki/daos/social/shl0ms-car) at 439 days, and it is the first row on the listthe one row whose clock is old because nobody opened the page. Every row aboveOn 20 September the list was headed by Getting started (/wiki/foundations/getting-started) and What is a caper (/wiki/foundations/what-is-a-caper) at 49 days; both were re-stamped on 21 September and left it. Every stamped row still on it was stamped and then edite…y is the one that gets re-run.

What the last-verified stamp h…a corrected population of 263, and again on 13, 15 and 17 September 2026 across the same 263, and again on 20 September 2026 across 270. The figures below are the 1720 September reading; where one …eached two-fifths. The oldest edit clock in the corpus stands at 369 days, the median at 4, and 2060 of the 26370 articles were edited within the past week, against 206 on 17 September, 209 on 15 September, 177 on 1…. The row holding that oldest edit clock is $CAR (SHL0MS) (/wiki…he thin part: 12 articles of 26370 carry one. For the other 2518 the last-edit date stands in.…t recent edit, by a median of 2531 days and a maximum of 469 – a median that was 25 on 17 September, 17 on 9 September and 3 a for… 37 on 11 September, 42 on 13 September, 44 on 15 September and 446 on 157 September, and a gap that wid…ery page the rotation touches. The maximum has changed hands as well as size: it sat on PsyDAO (/wiki/desci/ecosystem/psydao) until 17 September and sits on Getting started (/wiki/foundations/getting-started) and What is a caper (/wiki/foundations/what-is-a-caper) on 20 September, both stamped on 1 August 2026 and both edited in the eight hours before this reading. Four of them share one timest… most of this site's traffic. Read across the corpus on 20 September 2026, that effect orders the wiki. The six oldest freshness clocks here all sit on stamped pages: Getting started (/wiki/foundations/getting-started) , What is a caper (/wiki/foundations/what-is-a-caper) , PsyDAO (/wiki/desci/ecosystem/psydao) and AthenaDAO (/wiki/desci/ecosystem/athenadao) at 49 days, then DAO legal structures (/wiki/dao-governance/concepts/membership/dao-legal-structures) and AxonDAO (/wiki/desci/ecosystem/axondao) at 40. Two of the four at 49 days were edited in the eight hours before the reading. The seventh row is $CAR (SHL0MS) (/wiki/daos/social/shl0ms-car) at 39 days, and it is the first row on the list whose clock is old because nobody opened the page. Every row above it was stamped and then edited again. So the rule the section above … rather than as a measurement. The figure this page published for that clock was read off the wrong column, and the harness is where the error sat. Until 20 September 2026 the section above reported “the oldest clock in the corpus” from a script that read each page’s last-edit date alone, while the check that raises the notice reads the last-verified date wherever there is one and the edit date otherwise. On this corpus the two answer 39 days and 49 days. Nothing a reader saw was stale and no page was misdated; the number was not the number it was named. The harness reports both from this pass, labelled, and the paragraphs above say which is which. This page records the same failure once already, in the pass that corrected the population in the prose and left the script handing the old row back to the count – a rule written in prose and implemented in code is two copies of one rule, and the runnable copy is the one that gets re-run.

What the last-verified stamp h…lation of 263, and again on 13, 15 and 157 September 2026 across the sam…3. The figures below are the 157 September reading; where one …lock in the corpus stands at 346 days, the median at 24, and 2096 of the 263 articles were edited within the past week, up fromagainst 209 on 15 September, 177 on 13 September and 143 o…– carry one placed by hand on 3028 August 2026, after a governan… the mechanism they argue from, and rewritten on 30 August, 7 September and 13 September as that mechanism changed again. The clock has never raised t… of 25 days and a maximum of 446 – a median that was 17 on 9 S…, a maximum that was 37 on 11 September, 42 on 13 September and 424 on 135 September, and a gap that wid… rewritten on 22 August, on 7 and 12 September and again on 127 September, and – unless it is…ready reads updated: 2026-09-127 above last_verified: 2026-08-… rather than as a measurement.

What the last-verified stamp h…ation of 263, and again on 13 and 15 September 2026 across the same…3. The figures below are the 135 September reading; where one …lock in the corpus stands at 324 days, the median at 42, and 177209 of the 263 articles were edit…within the past week, up from 177 on 13 September and 143 on 11 September. The row h… of 25 days and a maximum of 424 – a median that was 17 on 9 S…, a maximum that was 37 on 11 September and 42 on 13 September, and a gap that wide… rather than as a measurement.

What the last-verified stamp h…oss a corrected population of 263, and again on 13 September 2026 across the same 263. The figures below are the 113 September reading; where one …lock in the corpus stands at 312 days, the median at 64, and 14377 of the 263 articles were edited within the past week, up from 143 on 11 September. The row holding that oldest …n of 25 days and a maximum of 3742 – a median that was 17 two days agoon 9 September and 3 a fortnight before that, and i maximum that was 37 on 11 September, and a gap that widens on every page the rot…ntially rewritten on 22 August, on 7 September and again on 712 September, and – unless it is…lready reads updated: 2026-09-0712 above last_verified: 2026-08-… rather than as a measurement.

What the last-verified stamp h…censuses read, and re-read on 911 September 2026 across a corre…63. The figures below are the 911 September reading; where one …is time it was in the corpus. That correction was made in the prose and not in the instrument, and the instrument is the part that gets re-run. These four sections are re-derived by script at the top of each policy audit rather than counted by hand – neutral point of view (/wiki/policy/neutral-point-of-view) says so, and it is the whole reason a figure here is expected to be current rather than merely dated. Read on 11 September 2026, that harness was still asking the database for every row and answering 264 articles, 32 of them on the Caper side, and 144 edited within the past week: the exact population these pages were corrected away from two days earlier. One of its four sections carried the guard and three did not, so the boundary agreed with the route table in the single place it happened to be written and nowhere else. This page had already set out what the excluded row does to these figures in particular – the sweep rewrites the log on every run, so it enters as an article edited today, every day – and the harness had gone on handing that same row back to the count each time it ran. It now takes the boundary from the route table itself rather than from a literal, and the figures below are the first on this page derived under it. Nothing here had gone stale, which is the part worth keeping: a rule written in prose and implemented in code is two copies of one rule, and correcting the readable copy leaves the runnable one to put the error back. The 180-day notice has never r…lock in the corpus stands at 391 days, the median at 76, and 1453 of the 263 articles were edited within the past week. The row holding that oldest clock is $CAR (SHL0MS) (/wiki/daos/social/shl0ms-car) , last edited on 11 August 2026, and it is also the single article neutral point of view (/wiki/policy/neutral-point-of-view) records as naming a caper without marking the mention. Two censuses built for different purposes landed on the same row, because they are measuring one thing from opposite ends: nobody has re-opened it. No page is stale by the publi… rather than as a measurement.

What the last-verified stamp h…censuses read, and re-read on 79 September 2026 across a corrected population of 263. The figures below are the 79 September reading; where one …hat it means, both are given. The denominator in all of this was one too large until this pass, and this page is where that is most embarrassing. The section above records that the maintenance log is written to a private tag path and has no public URL; the census below, and the ones on neutral point of view (/wiki/policy/neutral-point-of-view) and verifiability (/wiki/policy/verifiability) , were built by asking the database for every page belonging to no caper – which returns the log. So a page this wiki documents as unreachable was being counted as a public article three paragraphs later, and it is the worst possible row to have included here: the sweep rewrites it on every run, so it entered the freshness figures as an article edited today, every day. Re-probed on 9 September 2026 the log still returns a true 404 on its page, its markdown twin and its JSON endpoint. The reading below excludes it. The correction moves the 7 September population from 263 to 262 and today’s from 264 to 263, and it is the second time in a fortnight that a figure on these pages was falsified by the maintenance pass rather than by the world – the first time the pass changed the corpus, and this time it was in the corpus. The 180-day notice has never r…lock in the corpus stands at 379 days, the median at 57, and 202145 of the 263 articles were edit…t recent edit, by a median of 1725 days and a maximum of 37 – a gapmedian that was 317 two days ago and 213 a fortnight agobefore that, and hasit wideneds on every page the rotation has touched sinces. Four of them share one times… rather than as a measurement.

What the last-verified stamp h…the database on 23 August 2026, across the same 260 rows those censuses read (261 as of 1 September 2026). The 180-day notice has never rendered, and cannot yet. The oldest article here was created on 28 June 2026, which makes the whole wiki , and re-read on 7 September 2026 across 263. The figures below are the 7 September reading; where one has moved far enough to change what it means, both are given. The 180-day notice has never rendered, and cannot yet. The oldest article here was created on 28 June 2026, so the earliest date on which any page can go stale by the published rule is 25 December 2026 – and a page edited since then moves its own date out again. That is the honest form of the claim: an age in days is itself a fact with a shelf life of one day, and this sentence carried “56 days old –, under a third of its own threshold” until the same figure reached two-fifths. The oldest clock in the corpus stands at 357 days, the median at 5, and 148202 of the 2603 articles were edited within t…amp itself is the thin part: 112 articles of 2603 carry one. For the other 24951 the last-edit date stands in.…, and it also means that for 965% of this wiki last verified r… last_verified line. And all 112 stamps predate their page's m…t recent edit, by a median of 317 days and a maximum of 21. Five37 – a gap that was 3 and 21 a fortnight ago and has widened on every page the rotation has touched since. Four of them share one timestamp to the millisecond: four of the five policy pages, this one included, stamped in a single bulk write on 19 August 2026. This page is the exception, and it was the exception when the sentence above said five: it was re-stamped on 23 August 2026, in the pass that first wrote this section, and the sentence counted it in anyway. A bulk stamp records that a c…ntially rewritten on 22 August and again on 7 September, and – unless it is re-stampe… already reads updated: 2026-08-229-07 above last_verified: 2026-08-… rather than as a measurement.

The rotating audit Freshness i…ivate, so it has no public URL, – re-probed on 1 September 2026, the page, its markdown twin and its JSON endpoint each return a true HTTP 404 rather than a soft one – and what was checked, what wa…n the article's markdown twin.

What the last-verified stamp h…e 260 rows those censuses read (261 as of 1 September 2026). The 180-day notice has never… the rotation has not yet had. What a reader can see is a different thing wearing the same label. A May be outdated notice is also an editor's block, and three of this wiki's essays – The Exit Right (/blog/the-exit-right) , The Binding Vote (/blog/the-binding-vote) and The Personal Caper (/blog/the-personal-caper) – carry one placed by hand on 30 August 2026, after a governance change superseded the mechanism they argue from. The clock has never raised the notice; the rotation has, three times. The stamp itself is the thin … rather than as a measurement.

The rotating audit Freshness i…anently neglected. The log is public:not published. It is written to a tag path the route table marks private, so it has no public URL, and what was checked, what was corrected, and what is still outstanding are all readable to the maintainers rather than to a reader here. What a reader can see instead is the record each pass leaves on the article itself: the last-verified date shown here, and the updated and last_verified lines in the article's markdown twin.

What the last-verified stamp h…asier to write than to keep. This page was re-stampedrying it on this page, in the same pass that read the census and wrote this section – its claims re-checked against the freshness code, the stamping script and every external link in the policy category – which makes it, at the time of writing, the one art, turned up the reason it is not only a matter of discipline. The stamp is written by a script that updates the row, and the row's edit timestamp is maintained automatically on every write to that row, so the stamp lands a fraction of a second behind the edit date it is meant to postdate. A stamp can never be newer than its page's last edit: not 0 of 11 through neglect, but 0 by construction. And since the check prefers the stamp wherever one exists, it always reads the older of the two dates. Stamping cannot improve a page's freshness clock; it can only pin it – which narrows the advicle of the 260 whose stamp postdates its owabove into something more honest: stamp a page when you have genuinely re-checked it, and read the date it displays as a floor rather than last edi a measurement.

What the last-verified stamp h… easier to write than to keep. This page was re-stamped in the same pass that read the census and wrote this section – its claims re-checked against the freshness code, the stamping script and every external link in the policy category – which makes it, at the time of writing, the one article of the 260 whose stamp postdates its own last edit.

What the last-verified stamp has actually produced The rule above is checkable, so this section checks it, in the way neutral point of view (/wiki/policy/neutral-point-of-view) and notability (/wiki/policy/notability) check their own: by measuring the corpus rather than restating the rule. Read from the database on 23 August 2026, across the same 260 rows those censuses read. The 180-day notice has never rendered, and cannot yet. The oldest article here was created on 28 June 2026, which makes the whole wiki 56 days old – under a third of its own threshold. The oldest clock in the corpus stands at 35 days, the median at 5, and 148 of the 260 articles were edited within the past week. No page is stale by the published rule. What keeps this wiki current is the rotation, not the threshold; the threshold is a backstop against a failure the rotation has not yet had. The stamp itself is the thin part: 11 articles of 260 carry one. For the other 249 the last-edit date stands in. That is the fallback working as designed, and it also means that for 96% of this wiki last verified reports when someone last wrote rather than when someone last checked – the distinction the section above insists on. It is invisible to a reader, because a page with no stamp displays none: the markdown twin of an unstamped article (https://caper.network/wiki/daos/stablecoins/frax-finance.md) simply omits the last_verified line. And all 11 stamps predate their page's most recent edit, by a median of 3 days and a maximum of 21. Five of them share one timestamp to the millisecond: the five policy pages, this one included, stamped in a single bulk write on 19 August 2026. A bulk stamp records that a category was reviewed, not that every page in it was, and that is the weaker claim. The consequence sits in the clock, and it runs opposite to intuition. The freshness check reads the stamp where there is one and the edit date otherwise, so an unstamped page resets its own clock every time it is edited, while a stamped page's clock stops until someone stamps it again. Stamping a page you then keep editing therefore makes its freshness signal worse rather than better. PsyDAO (/wiki/desci/ecosystem/psydao) is the worked example: stamped 1 August 2026, substantially rewritten on 22 August, and – unless it is re-stamped – it will raise the May be outdated notice on 29 January 2027 however often it is revised in between. Its markdown twin already reads updated: 2026-08-22 above last_verified: 2026-08-01, which is how a pinned clock looks to the agents that make up most of this site's traffic. So the rule the section above states now has a measurement behind it: stamp a page only when you have genuinely re-checked it, and re-stamp after a substantive edit. This wiki's own record shows that is easier to write than to keep.

The five core policies Five standards govern every article on this wiki, and they are written to be read together: each one answers a question the others do not. - verifiability (/wiki/policy/verifiability) — sets the threshold for including material at all. - no original research (/wiki/policy/no-original-research) — limits what an editor may conclude from it. - neutral point of view (/wiki/policy/neutral-point-of-view) — governs how the material is framed. - notability (/wiki/policy/notability) — decides whether a subject earns an article of its own. - freshness — decides how long any of it stands without a re-check. They apply to articles about Caper and to articles about every other organisation the wiki covers, on the same terms.

The rotating audit Freshness i…nd records what it did on the wiki maintenance log (/wiki/operations/wiki-maintenance-log) . Because editing a page moves… outstanding are all readable.

Freshness is the Caper wiki's on-chain extension of verifiability. Facts about a fast-moving ecosystem decay quickly, so a claim sourced correctly a year ago may now be wrong. [1] (#ref-1) Factual pages should record when they were last checked and be re-verified against the live ledger.

Keeping pages current Set a Last Verified date when you confirm a page's facts, and add a May be outdated notice banner rather than leaving stale errors in place. Where a page states an on-chain value, prefer a derived block (timeline, grants ledger, funding graph) that re-resolves from the ledger [2] (#ref-2) — the maintenance sweep flags pages whose prose has drifted from the chain.

Freshness is the Caper wiki's on-chain extension of verifiability. Facts about a fast-moving ecosystem decay quickly, so a claim sourced correctly a year ago may now be wrong. [1] (#ref-1) A page is not finished when it is correct; it is finished when someone has checked that it is still correct, and recorded the date they checked.

The last-verified stamp Every article carries a last verified date, set when an editor confirms its facts against current sources and the live ledger. A page whose last verification is more than 180 days old renders a May be outdated notice at the top of the article, asking readers to treat its figures as unconfirmed and inviting a re-check. Where a page has never been verified, its last edit stands in for the stamp, so a page that is merely old is flagged the same way as one that is old and unchecked. Verification is not a re-read. Stamping a page means the editor went back to the sources: re-probed the routes the citations point at, re-derived the on-chain figures, and confirmed that a mechanism described in the present tense is still one anyone uses. Editing a page does not clear its stamp, so re-stamp after a substantive edit rather than assuming the edit counts.

Prefer a value that re-resolves The most durable fix for staleness is not to restate a number in prose at all. Where a page states an on-chain value, prefer a derived block (timeline, grants ledger, funding graph) that re-reads it from the ledger on every render, [2] (#ref-2) and where a protocol parameter is configurable, say so and point at where it is read rather than quoting a figure that a later governance vote will silently invalidate. A number frozen into a sentence is a claim with an expiry date; a number resolved at render time does not have one.

The rotating audit Freshness is maintained by a scheduled maintenance pass rather than by hope. Each run audits one category of the wiki, checks the external links in that subtree, refreshes the pages carrying the oldest verification dates, and records what it did on the wiki maintenance log (/wiki/operations/wiki-maintenance-log) . Because editing a page moves it to the back of the staleness queue, the rotation guarantees that every category is reached in turn and nothing is permanently neglected. The log is public: what was checked, what was corrected, and what is still outstanding are all readable.

Freshness is the Caper wiki's on-chain extension of verifiability. Facts about a fast-moving ecosystem decay quickly, so a claim sourced correctly a year ago may now be wrong. [1] (#ref-1) Factual pages should record when they were last checked and be re-verified against the live ledger.

Keeping pages current Set a Last Verified date when you confirm a page's facts, and add a May be outdated notice banner rather than leaving stale errors in place. Where a page states an on-chain value, prefer a derived block (timeline, grants ledger, funding graph) that re-resolves from the ledger [2] (#ref-2) — the maintenance sweep flags pages whose prose has drifted from the chain.