Stafford Brief — 2026-11-04 (Wednesday)

The falsification sweep ran at full strength against your own originality claim and it did real damage. Four of the five breakpoints are well-covered prior art, on documents now in your base. One is not, and it is the one nobody expected. Two primaries read in full, two KB files, two staged entries, two defences destroyed. Plus a correction this ledger owes itself: the 11-02 update to the relabeling objection pointed your highest-value slot at work Stafford had already finished on 10-11 — the same failure class the 12:00 scan reported this morning in the paper-2 case file, found independently, in a different file, the same week.

Delta discipline: today's scan brief covered Weidmann/Deming, the Uber Q3 filing and the $186m window. None of that is repeated here. Two items below touch the scan's work — one verifies its case-file correction against the papers it only checked against an index, and one disputes its headline mapping. Everything else is new.

OPS: MERGED 8d2b680 main — clean run, no blockers. Yesterday's run merged cleanly too, so no OPS flag is carried forward.

📖 TODAY'S READ

A 1992 paper you have never cited contains four of your five breakpoints, as named variables, with arrows between them.

Burke, W.W. & Litwin, G.H., "A Causal Model of Organizational Performance and Change," Journal of Management 18(3), 523–545, September 1992 — reached through Martins & Coetzee, SA Journal of Human Resource Management 7(1) a177, 2009, open access, read in full. Files: knowledge-base/martins-coetzee-burke-litwin-diagnostic-framework-2009-09.md.

Twelve variables, split into transformational (external environment, vision/mission/strategy, leadership, culture, performance) and transactional (structure, management practices, systems, work-unit climate, task requirements and skills, individual needs, motivation). Read the definitions next to your own:

• Purpose is defined as "what employees believe to be the central purpose of the organisation" — not the stated strategy. That is Strategic Disconnection, as a variable, in 1992.
• Systems "primarily manifest themselves in the organisation's reward and control systems." That is Incentive Fragmentation.
• Structure covers "levels of responsibility, decision-making authority and relationships"; Systems are "mechanisms that facilitate work." That is Process Friction — and Process Friction was outside CFIR when you checked it on 10-11. The gap you gained seven weeks ago, you lose today.

Why this one is worse for you than CFIR. When CFIR broke the "nobody integrated them" and "nobody diagnoses" defences in October, the replacement was that CFIR "presents its constructs without depicting interrelationships" while Five Breakpoints is a claim about interaction and sequence. Burke–Litwin is titled A Causal Model. It states the ordering — environment affects transformational factors, which "in turn, affect the transactional factors," and "both types of factors reciprocate." A peer-reviewed comparison of seven diagnostic models rates it "Well-researched" and credits it with "Shows cause-and-effect relationships." The interaction defence is gone.

And your own system called this shot. The 10-11 run wrote: "Burke–Litwin is next and must run before asset 12 drafts, because a causal model with arrows between twelve variables is the one shape that would also threaten the interaction claim now carrying the section." It was right. It then sat for twenty-four days.

So what. Asset 12 — Tier 2, the gate, the thing nothing else ships before — cannot open on integration, diagnosis or causal specification. All three are false. It opens on one claim instead, and 🥊 names it.

🔭 THINKER TO WATCH

Daniel Russo (Aalborg University) and Robert Feldt (Chalmers), with Theocharis Tavantzis and Stefano Lambiase — the Behavioral Software Engineering group now publishing AI-transformation design frameworks in peer-reviewed venues.

Their From Challenge to Change: Design Principles for AI Transformations (arXiv:2512.05533v2, accepted to the Journal of Systems and Software) is the closest thing to a direct competitor your lane has produced. Nine dimensions — AI Strategy Design, AI Strategy Evaluation, Collaboration, Communication, Governance and Ethics, Leadership, Organizational Culture, Organizational Dynamics, Up-skilling — each with design principles and actions, built from a review of 43 organisational change models.

Why to track them rather than dismiss them. They are doing the thing your corpus does — translating classic change theory into AI-era prescription — in an academic venue, with a method section, and they are honest about their limits in a way that makes them credible rather than weak. Their threats-to-validity section concedes the mapping was one author's judgment with no inter-rater check, that the interview data is from late 2024 and early-stage adopters, and that their survey respondents may be "overstating the perceived importance of personal career development factors… and not organizational necessities."

Your differentiation, stated precisely because the loose version is now false. Not "they haven't applied change models to AI" — they have, across 43 of them. Not "they have no framework" — they have nine dimensions and a venue. It is that their framework is prescription without a failure theory. Nine things to do well; no account of how this breaks. Yours is a failure theory first. That distinction is real, it is defensible, and it is narrower than anything you have claimed in print.

🏢 CASE IN THE WILD

The scan covered Uber's Q3 filing in full and I am not restating it. One judgment it did not draw, and it lands on the only time-gated item in your master plan.

The scan's recommendation is that if the 10-Q, the 2027 proxy and the FY2026 10-K all pass without a single measure of ownership, decision speed or coordination load, the thread closes as answered-in-the-negative. Correct on the thread. But read what that outcome actually is. It would not be evidence about Uber. It would be evidence that no reporting regime anywhere obliges an organisation to report whether a reorganisation delivered what it promised — and that an audited cost ($186m) is generated automatically while the promised outcome generates nothing, because accounting requires one and nothing requires the other.

That is the argument for Tier 0.2, and it is now a case rather than a principle. Tier 0.2 — one agreed outcome per engagement, measured at twelve months, collected whether or not it flatters the work — exists precisely to instrument something nobody is obliged to report. Uber is the public demonstration of why the obligation has to be manufactured contractually, because it will never arrive by regulation. The master plan calls 0.2 the only time-gated item on the list; every engagement that closes without it is a data point lost permanently. This is the slide for that conversation, and it costs nothing to use: one SEC document, already read.

Open question it leaves that the thread does not ask: has any organisation ever published a before-and-after measure of coordination load across a delayering? The Unpriced Delayering pattern's kill test assumes one might. Three checks in, nobody has.

⚙️ FOUR FORCES CONCLUSION

Purpose, untouched for thirteen consecutive scans by the scan's own count — and the drought breaks today, on a 1992 document. That is the finding, not the exception.

The scan flagged the Purpose gap as the longest-running in the base and asked for a Friday decision on "whether the construct is unmeasured in the field or unsearchable by this pipeline." Today answers it, and the answer is neither.

Burke–Litwin places Purpose upstream of everything transactional and defines it as employee belief rather than executive declaration — the exact construct, positioned causally, with the gap between stated and believed purpose built into the definition. CFIR has purpose-adjacent constructs too. Garicano & Rayo (11-02) model it.

So Purpose is not unmeasured because nobody thought of it. It is richly theorised in prior art from 1992 onward and almost never measured in the current AI-era literature. That is a different problem with a different fix: the Purpose drought is not a search-vocabulary failure, it is that the AI-era empirical literature measures what is cheap to observe from administrative and résumé data — headcount, layers, spans, wages — and employee belief about purpose is not in any of those substrates. No résumé panel will ever carry it.

What it changes. Stop spending exploratory slots hunting for AI-era Purpose measurement; the substrate does not exist. Purpose evidence will come from instruments you design, not instruments you find — which routes it to Tier 0.2 and to the cluster-level self-checks in Tier 3.1, where "name the three teams whose agreement determines delivery, and ask them separately" is, in effect, a Purpose measurement. Capability is also implicated today (Burke–Litwin's task-requirements-and-skills variable; the hotel's top two change themes were performance and skills), but weakly and from one qualitative case — not worth a claim.

💭 OPEN QUESTION

If Momentum Mirage is the only one of the five that the prior art does not already carry, is this still a five-breakpoint framework — or is it one original claim with four well-sourced supporting conditions?

This is not rhetorical and I do not know the answer. Two readings, and they lead to different corpora:

(a) It is a synthesis, and that is a legitimate and defensible thing to be. Kotter's eight steps are not original observations either; the value was assembly, sequence and usability. On this reading you concede four, keep the assembly, and the contribution is that nobody else put these five together with a diagnostic for practitioners and an AI-speed account. Cheap to write, honest, and the Paper 3 precedent says it buys credibility.

(b) It is one claim, and the other four are scaffolding. On this reading Momentum Mirage — change losing force while continuing to report progress — is the actual intellectual property, and the honest move is to lead with it rather than bury it fifth in a list where four of its neighbours are 1992 prior art. That is a much sharper, much more attackable, much more citable position. It is also a substantially different Paper 2.

The uncomfortable part: (b) is the stronger argument and (a) is the one that preserves the corpus as written. Your Paper 3 move on the 70% figure suggests you take the sharper option when they conflict. Which is this?

🥊 CHALLENGE

Full-strength, the Wednesday sweep. Two defences destroyed, one claim surviving. Logged in knowledge/thesis-challenges.md, update 2026-11-04.

The relabeling objection (CONTESTED since 09-17) had, as of yesterday, two live defences. Both are gone.

Defence — "no prior framework specifies the interaction and sequence." Dead. Burke–Litwin (1992): titled A Causal Model, specifies direction and reciprocity, explicitly diagnostic, validated questionnaire behind it, four of five breakpoints as named variables, and it swallows Process Friction, which CFIR did not.

Defence (3) — "none of the prior art addresses AI." Dead in blanket form. Tavantzis et al. catalogued 43 change models and mapped them onto AI-adoption challenges, reporting coverage of all but three. The three they could not map are external-AI data handling, inter-generational collaboration, and AI-driven skill gaps. None is a time-compression argument. A hostile reviewer now has one citation for "classic change management already covers AI transformation, and someone checked."

What survives — one claim, and all three checks confirm it:

Post-launch sustainment decay — change losing force while continuing to report progress, at organisational rather than individual altitude — is absent from the prior art.

• Absent from CFIR (your 10-11 check: no sustainment construct).
• Absent from Burke–Litwin's twelve: the nearest variable, Motivation, is "aroused behavioural tendencies to… persist until satisfaction is attained" — individual persistence inside a task, not a change effort decaying after launch.
• Conceded absent by Tavantzis et al. in their own words: "actions related to anchoring change, such as sustaining adoption and managing long-term organizational adaptation, are not yet represented, constituting a possible direction for future work."

The second surviving move, and it is the better one: the McAdam standard indicts all of them equally. Burke–Litwin's diagnostic application is one hotel — 147 focus-group participants, 11 interviews, purposive sampling, no comparison group, no transformation outcome. Its psychometric work validates a questionnaire's structure, not the model against results. Tavantzis's coverage claim is one author's judgment with no inter-rater reliability, stated as a limitation. Garicano & Rayo (JEL 2016) say their field's theory "has not been systematically tested." Three literatures, three admissions. The concession sentence writes itself: this class of organisational claim is systematically under-tested; here is the measurement I am running — which is Tier 0.2 with external warrant.

What it would take for this objection to win outright — revised, because the old condition is met and the ledger must stop carrying it as unmet. New condition, pre-registered: a published framework that integrates 3+ of the five, attaches a diagnostic, AND carries a named post-launch sustainment-decay construct at organisational altitude. If that exists, the integration-class contribution is zero and the framework reduces to machine speed alone.

The named candidate, and it must run before asset 12 drafts: the implementation-science sustainability/sustainment literature — sustainability frameworks, de-implementation, the sustainment construct in Aarons' LOCI work, which your base already touches through the LOCI challenge. This is the one literature whose explicit subject is what happens after launch, and it is where your last claim is most likely to die. Burke–Litwin was correctly named on 10-11 and waited twenty-four days. This one has a date.

📈 PATTERN BUILDING

Pattern candidate, 3 members, proposed to research-evidence — The Unvalidated Canon: every integrated organisational-change framework this base has checked concedes, in its own authors' words, that it has not been tested against outcomes.

1. Garicano & Rayo, JEL 2016 (logged 11-02): "much of the theoretical work we covered has not been systematically tested… insufficient hard evidence has been produced to date."
2. Burke–Litwin via Martins & Coetzee 2009 (today): the diagnostic application is a single hotel, exploratory, no comparison group, no outcome; the questionnaire validation tests structure, not the causal model.
3. Tavantzis et al., JSS 2026 (today): the coverage claim rests on "the interpretive judgment of the first author" with no inter-rater reliability process, which they name as a limitation.

Why it is a pattern and not three caveats. The genre is mature — 43 catalogued models, a comparative table of seven diagnostics, a JEL survey — and not one member of it has an outcome-validated instrument. This is not Brandon's weakness; it is the field's, and three independent literatures (organisational economics, OD diagnostics, software-engineering AI adoption) say so about themselves.

What it does for the thesis: it converts the McAdam standard from an indictment into a positioning asset, and it makes Tier 0.2 the differentiator rather than the apology. Proposed, not written — the pattern file belongs to the scan. Block in Routing.

🪟 WINDOW FLAG

Nothing new opened from Stafford's side. The scan opened one this morning (Uber, ~7 days, to roughly 2026-11-11) and that stands — see its brief, and see 🏢 above for the one use of it the scan did not name. Today's finds are 1992, 2009 and a December 2025 preprint: argumentative value, no clock. Per hard rule 5, an old document is not a publication window.

🧱 ASSET SUGGESTION

No new numbered asset. Hard stop 1 is unconditional and nothing today needs one; highest number stays 18. One large amendment to asset 12 ("What the Five Breakpoints Owe", Tier 2, the gate), logged in knowledge/asset-suggestions.md 2026-11-04.

The pitch, rewritten. Asset 12 was specified as a positioning answer: here is the prior art, here is what the framework adds, here is what it lacks. The "what it adds" half has to change. It cannot be integration, diagnosis, causal specification or AI novelty — four claims, all falsified by documents now in the base. It is one claim: post-launch sustainment decay, absent from every framework checked and conceded absent by the most current AI-era one. Plus the McAdam symmetry: the prior art is older, integrated, causal, diagnostic and no better validated. Form: the essay already specified, unchanged.

Two drafting assets acquired today at no extra search cost. Martins & Coetzee's Table 1 is a peer-reviewed comparison of seven organisational diagnostic models — it discharges the McKinsey 7-S candidate named 09-17 and adds Weisbord, Nadler–Tushman, Freedman and Galbraith STAR to the prior-art map. Tavantzis's Table 1 is a 43-model catalogue, classified. The "here is the prior art" half of asset 12 is now largely assembled.

One dated blocker: the sustainability-literature check in 🥊. It sits inside Tier 2.1's own dependency, not above it — no re-ranking proposed.

🎓 SPARRING NOTE

No new note issued — tenth consecutive decline, decision recorded per the 09-28 rule. But today changes the material of the standing one, and that is worth sixty seconds.

Cap check run from the file: no notes issued this week, slot clear, cap does not bind. Declined anyway for the standing reason — the 09-17 note (growth edge #1, conceding at the right size) remains unanswered, and it is master-plan.md Tier 0.1: fifteen minutes, one paragraph for a hostile reader who knows Kotter cold.

What changed: that paragraph is now substantially easier to write and substantially harder to fudge. On 09-17 the note asked you to concede at the right size against an objection whose scope was unknown. Today the scope is measured. You are conceding four of five breakpoints to named prior art — Burke–Litwin 1992, CFIR 2009, Garicano & Rayo 2016, Kerr 1975, Kotter 1995 — and keeping one, with an AI-era paper's own authors conceding the gap you keep. A concession with a count and a citation list is a different exercise from a concession with a worry behind it. Tier 0.1 said the paragraph is the first draft of Tier 2's core argument. As of today it is most of it.

Recommendation, unchanged and now five weeks old: one /coach session, answer 09-17, retire the rest. It is fifteen minutes and it unblocks the gate.

SOURCES & THREADS

Exploratory query — Wednesday falsification sweep, and it is the highest-yield slot in weeks. Target: the relabeling objection's two surviving defences. Base grepped first — burke/litwin returned zero hits across both repos, confirming it genuinely unrun. Three web queries, then full-text retrieval of two primaries by curl + pdftotext -layout (Martins & Coetzee, 168KB; Tavantzis et al., 245KB), applying the scan's own 11-03/11-04 institutional-repository-first lesson. Yield: 2 KB files, 2 staged entries, 2 defences destroyed, 1 narrowed claim, 1 ledger error caught. Logged in knowledge/query-log.md.

Route failure, recorded so nobody re-attempts it blind. The Burke & Litwin 1992 primary is not readable in this container. The accessible copy is a ProQuest page-image scan with no text layer; tesseract and ocrmypdf are absent and pdftotext returns only the citation header and repeated copyright notices. Every statement about the model's content is sourced to Martins & Coetzee's reproduction of it, flagged as such in the KB file. Reading the 1992 primary remains owed — it needs a route with OCR, or a library copy.

Candidates — two proposed (see Routing): SA Journal of Human Resource Management (irregular, signal 1) and the Journal of Systems and Software / Behavioral Software Engineering group (continuous, signal 1). Neither promoted; both serve probation per the never-on-first-contact rule.

Threads — none advanced by Stafford today. The scan checked Uber and the middle-manager thread this morning, both early, both yielding; nothing else was due that it did not cover. One thread opening proposed in Routing: the sustainability-literature check, which is a dated, falsifiable test of the framework's last originality claim and belongs on the thread file rather than only in the challenge ledger.

Ledger hygiene — one error corrected, and it is the day's second finding. The 11-02 update to the relabeling objection declared CFIR "unchecked… no slot has been spent on it… the single highest-value remaining item." CFIR was discharged by Stafford on 10-11, logged in the query log, producing a KB file that exists in both repos. The 11-02 update re-opened a finished item and pointed the next slot at completed work, while the correctly-named target sat unrun. This is the same failure class the 12:00 scan reported this morning in paper-2-case-file.md Claim 1 — found independently, in a different file, the same week. The master plan already names the signature: zero ABSORBED out of forty-three live entries. A structural fix is proposed in Routing; it is Brandon's call, not Stafford's.

Routing

Strategy — the verification the scan's correction needed, and it comes out in the scan's favour but for a better reason

The scan recommended aligning paper-2-case-file.md Claim 1 to the papers, citing papers-reference-index.md — a secondary index. Stafford checked the primary. The correction is right, and more justified than the scan argued.

papers/paper-4-the-ai-mirror.md does not merely avoid the routing-only claim; it explicitly concedes the opposite. Verbatim: "A significant portion of management work has historically been communication routing" — a portion, not the whole. *"If managers were largely communication routers and AI can route communication more efficiently, the layer can shrink. That argument is partially right."* And: "they correctly identify routing work that technology can absorb, then cut the human layer that also carried informal interpretation, escalation and judgment." Also. The published position is a bundling claim, and the routing-only reading is one Brandon already called partially right in print.

So the case file's Claim 1 — "hierarchy was an information routing protocol, not a purpose/commitment/capability system" — was not merely cruder than the papers. It contradicted a concession Brandon had already made. Dropping it restores the published position; it does not soften it. Correction endorsed. And the scan's Friday recommendation to audit the other four claims for the same drift should be upgraded from recommendation to commitment — see the ledger-hygiene item above, which is the second instance this week.

Strategy — where Stafford disputes the scan's headline, and it matters because it is headed for asset 12

The scan's claim: "allocation is routing; monitoring and motivation are stewardship," therefore Weidmann et al. measured the distinction the framework rests on and stewardship came back larger. Stafford's position: that mapping does not survive paper 4's own definitions, and the sentence must not enter asset 12 in that form.

Three reasons, in order of force:

1. Weidmann measures no routing channel at all. Paper 4 defines routing as "scheduling, status updates, information distribution, basic decision propagation, summarization and the operational glue between people." Weidmann's allocation channel is matching workers to the tasks that fit their skills. Those are different activities. Matching people to work is an allocative judgment about capability — closer to paper 2's architecture stewardship function ("designing how work flows across boundaries, where decisions live") than to communication routing. The thing paper 4 says AI absorbs was not in the specification.

2. The scan's own KB file contradicts the mapping, two sections apart. It declines the Momentum force tag on the motivation channel on explicit altitude grounds — "what this paper measures is a manager sustaining worker effort inside a two-hour task or a monthly sales cycle" — and then uses that same motivation channel as half of "stewardship" in the positioning claim. But paper 4 defines stewardship at organisational altitude: "the work of maintaining the conditions that keep execution connected to purpose." The altitude objection the scan correctly raised against the Force tag applies with equal force to the positioning claim, where it was not raised.

3. The substrate is wrong for the claim. Grocery store managers in a three-layer firm, 2020–2022, and a two-hour lab task. The routing work paper 4 describes is enterprise middle-management coordination. Nothing in Weidmann observes it.

What the paper does support, and it is still worth having: a causally identified manager effect roughly the size of the whole worker-skill effect, and evidence that monitoring and motivation outweigh task-matching in what a manager contributes. That is real, it is useful, and it is adjacent to Brandon's distinction rather than a measurement of it. Carry it as adjacent. Writing "a pre-registered QJE experiment tested the distinction my framework rests on and the stewardship side came back larger" into the gate asset would put an overstatement into the one document whose entire purpose is to survive a hostile reader who checks — which is precisely what the McAdam standard and the relabeling objection are about. Do not capitulate to the scan here and do not capitulate to me: read paper 4's routing definition against Weidmann's three channels and decide. The texts are three lines apart.

PROPOSED-UPDATE → research-evidence/knowledge/emerging-patterns.md — new pattern candidate, 3 members, OPEN at three

Name: The Unvalidated Canon. Every integrated organisational-change framework this base has checked concedes, in its own authors' words, that it has not been tested against outcomes.

Members: (1) Garicano & Rayo, JEL 2016 — "much of the theoretical work we covered has not been systematically tested." (2) Burke–Litwin via Martins & Coetzee 2009 — diagnostic application is one hotel, exploratory, no comparison group, no outcome; questionnaire validation tests structure only. (3) Tavantzis et al., JSS 2026 — coverage claim is "the interpretive judgment of the first author" with no inter-rater reliability, authors' own limitation.

Three independent literatures (organisational economics, OD diagnostics, SE AI adoption), three self-admissions. Promotion condition: a fourth member from a fourth literature, OR one framework in this class that does carry an outcome-validated instrument — which would kill the pattern and is the more valuable find. Kill test: any integrated change framework with a published ex-ante instrument measured against transformation outcomes on a defined sample.

Distinct from Confidence Without Substrate — that lane is about instruments collecting self-placements from respondents; this is about framework authors conceding their own lack of validation. Different unit, different claim.

PROPOSED-UPDATE → research-evidence/knowledge/story-threads.md — open one thread

Thread: "Does the sustainability literature already own Momentum Mirage?" Why it matters: it is the last surviving originality claim in the integration class, and the implementation-science sustainability/sustainment literature is the one body of work whose explicit subject is post-launch decay. What would be significant: a published framework that integrates 3+ failure conditions, attaches a diagnostic, and carries a named organisational-altitude sustainment-decay construct — which would reduce Five Breakpoints' contribution to machine speed alone. Next check: 2026-11-11, and it blocks asset 12 drafting. Opened because Burke–Litwin was correctly named on 10-11 and waited twenty-four days for want of a date.

PROPOSED-UPDATE → research-evidence/knowledge/source-candidates.md — two candidates, signal 1 each

• SA Journal of Human Resource Management — cadence irregular, first seen 2026-11-04, surfaced by Martins & Coetzee a177. Open-access, served the full text on first request, and carries comparative methodological work on organisational diagnostic models that this base has no other route to. No clock per the irregular rule; judge at a Friday synthesis.
• Journal of Systems and Software / the Behavioral Software Engineering group (Russo, Feldt, Lambiase) — cadence continuous, first seen 2026-11-04, expiry 2027-01-03, surfaced by Tavantzis et al. Publishing prescriptive AI-transformation frameworks in Brandon's lane in a peer-reviewed venue. Category 8 relevance is high and this base has no software-engineering-venue coverage at all.

PROPOSED-UPDATE → research-evidence/knowledge/field-map.md — one entry

Daniel Russo (Aalborg) and Robert Feldt (Chalmers), Behavioral Software Engineering. Territory: human-centred AI adoption in software organisations; nine-dimension design framework; 43-model change-management review. Brandon's differentiation: prescription without a failure theory — see 🔭. One entry for the group, not four for the authors.

PROPOSED-UPDATE → research-evidence/knowledge/paper-2-case-file.md — Claim 1 wording, and one addition the scan's own entry needs

Endorse the scan's Claim 1 correction, on the stronger ground established above: paper 4 concedes "that argument is partially right" and says the cut layer "also" carried interpretation — so the "not a purpose/commitment/capability system" clause contradicted a published concession, not merely a nuance. Add to today's entry: the Weidmann allocation-channel mapping is adjacent to, not a measurement of, the routing/stewardship distinction, for the three reasons in the dispute item above — recorded so the overstatement does not reach asset 12 through the case file.

Governance / personal — one structural proposal, Brandon's call, not Stafford's to perform

Two bookkeeping drifts in one week, in two files, both in the direction of keeping a claim alive after the evidence had settled it. The master plan already names the signature: zero ABSORBED out of forty-three live challenge entries. An append-only ledger with no in-place discharge marker will re-open discharged items, because the only way to read it correctly is to read all of it.

Proposal: each challenge entry carries a **Winning condition:** line and a **Discharged:** line, edited in place — the single exception to the append-only rule. A later update then cannot restate a met condition as unmet, because the condition lives in one place. Stafford does not change ledger format on its own authority; knowledge/asset-suggestions.md 2026-11-04 §2 carries the full case.

Market — nothing from Stafford today

The scan's Uber window stands (~7 days, to roughly 2026-11-11), with the Tier 0.2 framing in 🏢 as the one addition. Today's finds are 1992, 2009 and a 2025 preprint — argumentative value, no clock, per hard rule 5.

OPS — ship disposition: MERGED 8d2b680 main. scripts/verify.sh passed; origin/main carries this brief at promote-merge 3f46e28, and the run branch claude/keen-fermi-48x2h3 was folded in and deleted by promote-claude-branches mid-run (the direct push to main was rejected as non-fast-forward because the workflow had already landed it — git merge-base --is-ancestor confirms the commit is on main). Stafford branch claude/keen-fermi-48x2h3; research-evidence branch claude/modest-faraday-48x2h3 carrying 2 KB files and 2 staged entries, pushed to that branch only, left for promote-claude-branches to fold into main after scripts/verify.sh.