The stowaway curriculum — forest school as the medium for the untested-but-mandatory

Published live (2026-07-02). The third definite treatment, and the first brought in from the field: the spec was developed with Max, a design-technology educator with children at a primary near Walton-on-Thames, Surrey — the treatment's centre of gravity is his observation, not ours. Process honesty, printed before anything else: this run's evidence fleet never launched. The worker-dispatch surface and every live web tool were refused at the permissions wall, so the run degraded to a single agent's pass over its trained index (current to January 2026) plus the two humans' field knowledge. No live source was consulted. The structural analysis — permission surfaces, rung-by-rung rationality — does not need ground truth the way the evidence legs do, and the verdict is split accordingly: passes-on-paper, the weakest verdict the gate has ever issued, and deliberately so. Condition zero is re-running every evidence leg against the live world. A treatment that hid this would be worthless; one that prints it is still worth publishing, because the design logic either survives contact with the evidence or it doesn't, and now the contact points are listed.

Update, 2026-07-03: condition zero is discharged. The fleet launched the next day and ran every leg against the live web — see the Evidence pass section below. The verdict has climbed accordingly: passes-on-paper → passes-with-conditions. The paragraph above stands as the historical record of the degraded run.

The spec as received

Three field observations, taken as premises and then attacked:

1. The medium premise. Outdoor learning is not a Wednesday-afternoon enrichment bolt-on; it is a medium through which much of the primary curriculum can be delivered, often better. Most primaries already have a forest-school patch or equivalent wild corner — the infrastructure exists and sits nearly idle, which makes it this domain's instance of held-value-not-reachable: a frozen asset in the school's own grounds, used an hour a week if at all. Risk assessments are routine admin, not a wall.
2. Max's scarcity finding. The binding scarcity is teacher time. Teachers are pressured to cover too many areas, so they ration coverage — fast, surface, worksheet-shaped — rather than teach deep, holistic, cross-subject. The queue signal in this domain is not a waitlist for forest school; it is teachers queueing for better means to mandated ends.
3. The accountability constraint. KS2 SATs test English and maths, full stop — the science tests were abolished in 2009 and science has been teacher-assessed since. Everything else on the national curriculum is mandatory but untested at scale. Any design whose pitch is "deeper understanding, but SATs might dip" fails at the median instantly: no data exists either way, schools deny drilling, and no head volunteers to run the experiment.

The narrowing — and what attacking it did to it

The design as specified: route around the constraint entirely. First rung = forest delivery of the non-tested-but-mandatory curriculum (science, geography, art, DT) — zero SATs surface, mandatory-coverage relief, existing patch, routine risk assessment, one teacher can start tomorrow. Literacy enters as a stowaway (the forest session is the writing stimulus); numeracy last, once the channel is normal; no event at which anyone approves "changing how we teach maths". Pitch shape: time-efficiency, never depth-versus-results.

The run's job was to attack the narrowing, not admire it. Four attacks landed and changed it:

• The stowaway had to be re-specified as content, not timetable. "Write about what we found" is only zero-SATs-risk if the forest session never claims an English slot — it feeds one. The English lesson stays exactly where it is on the timetable, delivered exactly as before; what changes is that its stimulus arrives from Tuesday afternoon's pond instead of a laminated prompt. The moment the design says "forest literacy replaces the literacy lesson", the SATs risk it routed around walks back in the front door. Stowaway means stowaway: the cargo doesn't book a seat.
• The time-efficiency pitch met the wellies overhead. Getting thirty children booted, out, worked, back and changed has a fixed transition cost that can eat a third of a single period. The pitch survives only in block-scheduled afternoons (which most primaries already run for foundation subjects) and only because the double-count pays the overhead: one afternoon covers the science objectives and generates the week's writing material and replaces the engagement-manufacturing planning the teacher was doing at 9pm. But the honest arithmetic — minutes out, objectives covered, planning minutes saved — has never been measured by this run and is a named condition. If the true accounting is "same coverage, more minutes, happier children", the design's only remaining pitch at the median is the one the spec forbids.
• The video objection. Any rich experience generates writing stimulus; a wildlife documentary does it with no wellies. Double-counting alone doesn't select the forest. What selects it is narrower and should be claimed narrowly: the science objectives whose objects of study are physically in the patch — habitats, classification, plants, seasonal change, rocks and soils, materials in weather — where the worksheet is literally a picture of the thing forty metres from the classroom. A mapping that claims everything teaches better outdoors is advocacy; one that concedes the indoor-dominant objectives (and marks them) is a tool. The artefact spec inherits this: honest concessions are a feature, load-bearing.
• The confidence wall — the attack that survived. Risk assessment is admin; solo outdoor behaviour management is not. Taking your own class of thirty out alone, weekly, is a real skill and a real fear, and no mapped scheme dissolves it. Rung 1 survives by self-selection: the adopting unit is one willing teacher per school, not the median teacher — and the design genuinely only needs one to start. But that purchase relocates the median test to the spread rung (rung 4, below), and the run could not price what carries it. Printed as residue, not managed away.

The ladder — the stowaway ratchet

Six rungs and a rung 0. Every rung individually rational on day-one terms for its adopter; no rung requires the next; no event at which anyone approves changing anything tested.

Rung 0 — the mapped patch. The battery: a season-sequenced library of outdoor sessions mapped to national-curriculum objectives at objective granularity, for the untested-but-mandatory subjects — with the indoor-dominant objectives conceded in print, and each session at photocopier-parity planning cost (pick up, walk out; if it needs an evening's adaptation it loses to the worksheet). Per the artefact hunt below, rung 0 is mostly curation and legibility, not authorship — fragments exist; what doesn't exist is the canonical, free, sequenced thing a head can hand an inspector.

Rung 1 — one teacher, one afternoon. The self-selected teacher takes an existing foundation-subject afternoon into the school's existing patch using the mapped sessions. Own class, own timetable slot, standing parental consent, the same risk-assessment template the school already uses for the playground. Coverage evidence identical or better; planning cost lower from week one. Permission events: zero.

Rung 2 — the stowaway. The session's material becomes the week's writing stimulus — recount, report, instructions, poetry, all of it framework-legible for teacher-assessed writing. The English lesson doesn't move; it gets better cargo. Nobody approves anything; the literacy coordinator notices the writing improved before noticing where it came from.

Rung 3 — the head's story. The teacher's coverage map is, for free, the head's answer to the foundation-subject deep dive: broad-and-balanced, first-hand experience, curriculum intent with mud on it. The head starts spending the story before adopting a position on the pedagogy. This rung matters because it converts the design from tolerated to cited without a decision event.

Rung 4 — the colleagues, and the re-priced PPA hour. Spread by observed workload relief: the year-4 teacher sees the year-3 teacher's planning folder thin out. And the one pot of existing money in the domain gets re-priced rather than newly spent — schools already pay external forest-school leaders to cover PPA time as enrichment childcare; handed the mapped scheme, the same purchased hour delivers mandated coverage. No new spend, no new permission; an existing invoice changes what it buys. This is the rung the whole design's scale lives on, and the run's sharpest unanswered objection lives here too (below).

Rung 5 — numeracy, last, as a free option. Maths outdoors only where the forest is genuinely dominant (measurement, scale, data collected from the real world) — explicitly optional, never the pitch, possibly never fired. Like the AV endgame in the mobility treatment, the ladder must pay without it, and does.

Rung 6 — the cluster precipitates, unforced. Practising teachers find each other; moderation of forest-generated writing across schools; a TeachMeet; possibly a MAT-level scheme. Nothing legal forces a wrapper here — see the candidates section, because that absence turned out to be a finding about institutions-precipitate, not about this design.

The mapping-artefact hunt — both branches fire at once

(Provenance: trained index to January 2026, unverified live. This is the run's centrepiece question and its least-verified section — condition zero applies here first. Discharged 2026-07-03: the hunt was re-run live; the section's shape survived, with corrections — see the evidence pass below.)

The discovery question was binary: if a curriculum-mapped forest scheme exists and two motivated, well-connected people had never heard of it → legibility gap; if it doesn't exist at objective granularity → the mapping is the battery. The trained index says: both, simultaneously — the artefact half-exists, in fragments, and every fragment is illegible from where Tom and Max stand.

The fragments, as the index has them: Learning through Landscapes (the UK's outdoor-learning charity) publishes free curriculum-linked outdoor lesson resources by subject and key stage; Twinkl — which most primary teachers touch weekly — carries a large outdoor/forest-school category tagged to curriculum aims; commercial subscription schemes exist in the shape of The Muddy Puddle Teacher and similar; Bloomsbury publishes a book series in the shape of The National Curriculum Outdoors (per-key-stage volumes mapping outdoor lessons to objectives); the RHS Campaign for School Gardening and the DfE's own National Education Nature Park carry curriculum-linked material; the LOtC Quality Badge accredits venues and providers but maps nothing; OPAL does playtimes, not curriculum. And the Forest School Association — the body that owns the brand — definitionally refuses the mapping: its six principles define forest school as a long-term, learner-led, play-centred process, explicitly not curriculum delivery. The refusal is principled and self-aware, and it means the movement with the name will not build this bridge; "curriculum-mapped forest school" is, to the brand's own gatekeeper, a category error.

So what's missing is not authorship but canonisation: there is no White Rose of the outdoors — no free, complete, season-sequenced, objective-mapped, Ofsted-citable scheme with enough brand gravity that a head can name it in a deep dive and an inspector nods. White Rose is the existence proof that in this exact population (time-poor primary teachers) a mapped scheme with those properties propagates teacher-to-teacher without a campaign — the bridge-shaped object in this domain is known to work; one has simply never been built pointing at the school's own patch. Rung 0 is therefore curation, sequencing, and legibility-engineering over mostly-existing material — cheaper than authorship, and the 2028 curriculum reset (below) says exactly when to ship it.

One more local fact, almost comic (verify on foot — it's walking distance): Walton-on-Thames already contains the existence proof. Ashley C of E Primary School spent years delivering a nature-principled curriculum (the Harmony work associated with its former head), mainstream state primary, SATs intact, results reportedly fine. It proves deliverability and results-compatibility in this exact town — and its adoption path (a visionary head, a whole-school redesign) is precisely what the bridge test forbids, which is presumably why it did not replicate down the road. The narrowed design is the bridge-shaped retry of what Ashley proved deliverable: same medium, no hero required. That the existence proof sat two miles from Max's school and never propagated is the legibility-gap finding in its purest form.

The gate — passes-on-paper

(Superseded 2026-07-03: condition zero discharged, verdict now passes-with-conditions — see the evidence pass below. This section stands as the record of the paper-only verdict and of the conditions as first issued.)

The gate could not run its full adversarial fleet (same permissions wall), so what follows is one adversary's honest pass, not three independent verdicts — another reason the verdict is the weak form. Passes-on-paper: the structure survives the bridge test as specified — every rung individually rational, permission surface near zero, the tested core untouched by construction — and the evidence legs are open. Conditions before any part ships as "ready":

0. Re-run the evidence legs live. The artefact hunt (does the White-Rose-shaped gap really exist, or did the index miss the incumbent?), the Ashley/Harmony facts, the FSA position, the 2028 dates. Everything below inherits this.
1. Measure the wellies arithmetic in one real classroom: transition minutes, objectives covered per afternoon, planning minutes saved. If the double-count doesn't pay the overhead, the pitch reverts to depth-versus-results and the design fails its own spec.
2. Price the confidence wall. What fraction of teachers will solo-take their class out weekly, and what does the willing-teacher-per-school density actually look like? (The Natural Connections project — the largest English study of school outdoor learning — found time, curriculum pressure and confidence the top barriers, and teacher-reported engagement gains near-universal; re-read it with rung-4 eyes.)
3. No attainment claims, ever. The EEF's outdoor-adventure-learning strand was withdrawn for lack of evidential security; there is no attainment shield to hide behind, which is fine, because the design never pitches attainment. The discipline must hold under marketing pressure.
4. Verify the PPA re-pricing rung against a real invoice: can an existing enrichment-cover contract legally and practically deliver mandated-coverage sessions, and will a head sign that sentence?
5. The artefact must ship with its concessions printed — the indoor-dominant objectives marked as such — or it is advocacy and fails the video objection.

The walls ledger

What this run gave the four candidates

The run obligations, discharged explicitly:

bridge-catchment — the catchment drew itself sharply. Inside: the self-selected confident teacher in a green-patched school — and the run had to face the gerrymander objection head-on: redefining the adopter unit ("one willing teacher per school, not the median teacher") is legitimate only because one teacher genuinely suffices to start the ratchet; the discipline is that the unit must be whoever must act for the value to arrive, and the median test then re-applies at every later rung rather than disappearing. Outside: tarmac-only schools (physical), high-turnover and supply-heavy schools (the ratchet needs a persistent teacher to accumulate normalisation), and — the exact shape the energy treatment found in fuel poverty — schools in special measures, which are simultaneously the most curriculum-narrowed (most in need of exactly this) and the least able to tolerate any deviation. The remainder is named but thinly priced: nothing in this design carries the special-measures cohort; only the policy layer can, and the design deliberately asks nothing of it. Priced-not-moralised: half-met, and printed as half-met.

institutions-precipitate — this domain delivered both a genuine non-energy instance and a new wrinkle. The instance: UK forest school itself precipitated — practice first (a 1993 college trip to Denmark, then sessions), a training market crystallising out of the practice, and the Forest School Association arriving in 2012, nineteen years after the first session, as paperwork catching up with an existing practitioner community. No founding act, no rails, no legal forcing. But that last fact is the refinement: where the second-commodity run found precipitation requires rails when the law forces a wrapper, this domain shows unforced precipitation — possible, slow, and optional, because nothing legally requires teachers' practice to incorporate. The cost of unforced is fragility: what precipitates without pressure stays thin. And the wrinkle worth its own line: the precipitate can occlude — the FSA, once crystallised, acquired an identity interest and now guards the definition against the widest catchment (its principled refusal of curriculum mapping is precisely what keeps this bridge unbuilt). Institutions precipitate; then the precipitate has interests. The counterexample hunt, honestly run, also found the anti-case in-domain: the national curriculum itself (1988) is a pure founding act that worked — consistent with the candidate's scope condition, because a national standard is exactly the thing whose first unit of value requires the whole.

find-the-waitlist — Max's twist tested, and it half-holds. The queue is real and enormous: teachers pay for Twinkl subscriptions (often out of their own pockets — the strongest behavioural signal in the domain), schools buy Kapow/PlanBee/Cornerstones-shaped foundation schemes, hire enrichment providers to cover PPA, and spend on trips — all of it unsubsidised, unexhorted spending on making mandated content deliverable in the time available. That is a true positive for the generalised lens: a queue for better means to mandated ends. But the run must log the precision finding: the queue is for the function (planning-cost relief), not for this instance of it (the forest medium). Teachers queueing at the photocopier are not queueing for wellies. The lens upgrade: identify what the queue is for at the level the queuers experience it — a bridge that enters through any other door than the queue's own function will stand unused beside a visible queue. Hence the photocopier-parity spec on rung 0: the artefact wins only inside the function the queue is actually for.

the-closing-window — the most surprising result: this domain's clock runs backwards. There is no enclosure race on the asset (nobody is buying the patches; the slow MAT-standardisation enclosure is real but decades-grained). Instead there is a scheduled opening: the national curriculum is being rewritten now (final review report late 2025; new curriculum slated for publication around 2027, first teaching from September 2028 — dates to verify, condition zero). On that day every scheme of work in England goes stale simultaneously and every school re-plans at once — the one moment a forest-mapped scheme competes against incumbents on equal footing rather than against sunk adoption, and the largest legibility window this domain will see in a generation. So the candidate gains a new type: alongside the race (mobility) and the revocable exemption (energy), a window that opens on a known date — and the doctrine extends from "move before it shuts" to "have the artefact ready the day it opens". Authoring an objective-level mapping against the 2014 curriculum today buys a two-year shelf life; building the practice channel (normalised outdoor delivery, kit, consent, one teacher per school) buys something curriculum-version-independent. Sequence accordingly: practice now, canonical artefact timed to the reset. A domain without a closing window would also have been a valuable negative; this is better — a domain with a starting gun.

The sharpest objection the run could not answer

Rung 4 has never been observed. Anywhere. The energy treatment had Bethesda already standing mid-ladder; the mobility treatment at least had Liftshare's repeat-pairing data for its load-bearing rung. This ladder's entire scale story — colleagues adopting by observed workload relief, the PPA hour re-pricing itself — is, on this run's evidence, pure design logic with zero observed instances of teacher-to-teacher lateral spread of curriculum-mapped forest delivery. The one local existence proof (Ashley) spread by hero-head, which is the wrong mechanism, and demonstrably didn't spread between schools at all. Everything hard in this design was pushed to rung 4 — the confidence wall, the median teacher, the honest arithmetic — and rung 4 is exactly where the run has nothing. If lateral spread doesn't happen, the design is not wrong, but it is small: a permanent scattering of rung-1 teachers, each an island, which is roughly what outdoor learning in England already is. The whole treatment climbs or stalls on whether anyone has ever seen rung 4 fire.

Evidence pass (2026-07-03) — the fleet launches

One day after publication, condition zero was discharged: the fleet that was refused at the permissions wall launched against the live web — four legs (the artefact hunt, the queue, the Walton ground, the evidence base and the window), every figure below from a source fetched 2026-07-03. Process honesty runs both directions: the tools worked this time; two primary pages resisted direct fetch (the EEF strand page returns 403 to non-browsers; one candidate scheme's domain no longer resolves) and are flagged inline where they matter. What follows is what the live world did to the trained index — confirmations, corrections printed as corrections, and two facts the index could not have had.

The artefact hunt, live — verdict (c): the mapping is the battery

Nothing found is simultaneously free + complete + objective-granular + season-sequenced + canonical. The "half-exists in fragments" hedge is confirmed and sharpened:

• Bloomsbury's The National Curriculum Outdoors is the closest fragment, verified in detail: five volumes (KS1, then Y3–Y6), ten subjects each, objectives named per lesson — but six lessons per subject per year. A sample, not coverage; paid, with a free companion layer of curriculum maps on top; no canonicity claims; still expanding (a 2024 climate volume) — a live market, not a saturated one.
• The one unresolved candidate: Outdoor Learning Made Easy's "Pioneer Scheme" — on its indexed marketing the closest shape to rung 0 (ten levels, 240 sessions plus 1,500 resources, "full curriculum links for every activity", assessment trackers). Its domain failed DNS from two independent runs, so it is characterised from indexed copies only: paid subscription, and the spine is its own eight-principle progression with curriculum links attached — the inverse of the rung-0 spec, which puts NC objectives at the skeleton. Flagged unverified-at-source.
• The counter-evidence find, carried honestly: Educate Outside ships a full-year outdoor scheme and prints, as a feature: "This scheme isn't designed to map perfectly to curriculum checklists." Someone built the adjacent product and chose looseness. Two readings, both kept: objective-granularity mapping has been tried and found commercially unattractive; or it has never been tried in the one combination White Rose proves propagates in this population — free core, complete coverage, brand gravity. Rung 0 stands on the second reading, and now says so out loud.
• The rest of the fragment map verified as fragments: Learning through Landscapes is freemium at topic granularity; Twinkl's outdoor category is aims-level and says so ("align, but are not bound to"); The Muddy Puddle Teacher is mostly CPD, not a lesson library; the LOtC badge accredits providers and maps nothing; OPAL is playtimes; the DfE's Nature Park is free and real but supplementary, science/geography/climate-shaped. And the big three foundation-scheme vendors — Kapow, PlanBee, Cornerstones — have no outdoor-delivery strand at all, confirmed live: the vendors who already own the objective-mapping muscle have not pointed it at the patch.

Correction — the FSA claim, softened. The note said the Forest School Association "definitionally refuses" curriculum mapping and calls it a category error. The live pass verified the six principles verbatim (long-term, nature, risk, holistic, leadership, community; play-centred, learner-led) — but found no explicit anti-mapping statement on the FSA's published principles or "What is Forest School?" pages. Curriculum-mapped delivery clearly sits outside what the principles define as forest school, and the ethos-tension is real; the "refusal" was an inference dressed as a quote. Softened accordingly — and the occlusion wrinkle this note fed into institutions-precipitate weakens with it, from verified-refusal to inference-from-ethos. Logged here for the next register reconciliation.

The queue, priced

• The PPA-cover market is real, commodified, and priced — the strongest verification in the pass. Multiple providers market forest-school PPA cover region by region at £150–£300/day (general PPA cover £150–£250/day), weekly cadence, per class. The invoice rung 4 proposes to re-price exists as a standing commercial category.
• White Rose verified as the existence proof, with a nuance: ~70% of schools follow it; the free core is real. But a paid premium layer has since grown around the free core — even the existence proof monetised its periphery, which is itself the queue thesis restated.
• Twinkl: £7.99–£8.49/month individual tier; turnover £66.6m (FY to April 2023). No Twinkl-named survey of personally-paid subscriptions was found; the adjacent evidence is strong — a Tes survey of 1,800+ teachers found 94% pay for school essentials out of pocket, and a cost-of-living survey put education staff spending their own earnings on classroom supplies above 50%. The note's "often out of their own pockets" is downgraded from asserted fact to consistent-with-adjacent-survey-data.
• School contributions to visits and trips: £200.9m in 2023–24 (LA-maintained schools only, +8.3% year on year); no primary-only or per-pupil split is published. Foundation-scheme spend verified at the price-list level: Kapow £215–£333+VAT/yr by school size; Cornerstones £500–£2,800/yr. Real, unexhorted money for making mandated content deliverable — the queue holds.

The Walton ground

• Ashley verified — and darker than the note had it. The Harmony curriculum was real (half-termly enquiries on nature's principles), the results claim survives scrutiny (top 16% nationally by KS2 attainment across 2016–19 and 2022), the school went Good → Outstanding in 2007 under Richard Dunne, head for eighteen years. Then the live record: Dunne was forced out in autumn 2019 by the academy trust amid performance allegations — all of which the Teaching Regulation Agency dismissed in 2020 as not amounting to potentially serious misconduct even at their highest. Ofsted: Outstanding → Requires Improvement (Nov 2021) → Good (Mar 2024). The school's current website carries no mention of Harmony, Dunne, or nature's principles anywhere. The practice didn't merely fail to replicate down the road; it failed to survive in the school that invented it. Dunne's Harmony Project now works with schools nationally — no evidence of any Walton-area school adopting.
• Local provision confirmed dense: private forest-school providers operate in Walton itself and the surrounding villages (Rewilding Forest School, Outdoor Owls, West End Playgroup, Piccolo, Walton Montessori). Oversubscription could not be verified or refuted online — providers don't publish waitlists; still an on-foot item, as the note predicted.
• Surrey runs a county outdoor-learning service (SOLD, four residential/day centres) — enrichment-and-residential shaped, not curriculum-delivery shaped: the incumbent frame, institutionalised at county level.
• Not discharged: the site-by-site green-space survey of Walton primaries. The walk stays on the list.

The evidence base, and the window

• Correction — the EEF strand was not withdrawn. It is live in the toolkit, rated "very low" evidence security (~6 studies, none independently evaluated), with no months-progress figure displayed and an explicit note that this is absence of secure evidence, not evidence of absence. The substance the note built on — there is no attainment shield — survives; the word "withdrawn" does not. (Direct fetch 403'd; corroborated from indexed copies — flagged.)
• Natural Connections verified: 125 schools, 40,000+ pupils, final report July 2016; teacher confidence and fragmented provision the headline barriers, with time and curriculum pressure underneath; over 90% of schools agreed outdoor learning was useful for curriculum delivery. The confidence wall is exactly where the note put it.
• Forest Research's corpus is as characterised: wellbeing, confidence, social development — small qualitative samples, no attainment claims.
• The attainment vacuum is confirmed by hard live search: no completed RCT or quasi-experimental UK study of curriculum-embedded outdoor delivery and attainment, 2020–2026. One in-flight cluster-RCT exists — outdoor versus indoor science teaching, primary, started September 2025, estimated completion late 2026 — the first external event that could test the no-attainment-claims discipline from either direction. A new clock item: when it reports, the discipline gets audited by the world.
• Ofsted: outdoor learning is "an essential element of a broad and balanced curriculum" — rung 3's mechanism verified. But the late-2025 inspection framework removed named subject deep dives from ungraded inspections (lines of enquiry and learning walks instead), so rung 3's "deep dive" wording is dated; the head's-story mechanism survives in a softer, less subject-addressable form.
• The window, verified to the day, and stronger than the index knew. Francis Review final report: 5 November 2025. New national curriculum: published spring 2027. First teaching: September 2028, single unphased launch — every scheme of work in England goes stale on one date, as the note claimed. Two strengtheners the trained index could not have had: the new curriculum will be digital and machine-readable, which makes an objective-level mapping cheaper to build and to verify at exactly the moment the incumbents reset; and KS2 SATs stay, with only minor refinements (an amended grammar test) — the accountability constraint the whole design routes around is stable through the window. The clock re-verifies as a scheduled opening, now with a sharper edge and a published gun date.

The verdict, revised — passes-with-conditions

Condition zero is discharged and the structure survived contact: the artefact gap is real, the queue is real and priced, the window is real and dated, the attainment vacuum is real and now bounded by an in-flight trial. Per the treatment spec's grading, the verdict climbs from the gate's weakest form to the energy run's form — passes-with-conditions — and the remaining conditions are now purely field-shaped; nothing left on this list can be discharged from a desk:

1. The wellies arithmetic (unchanged): one real classroom, transition minutes and objectives counted.
2. Rung 4 (unchanged, and untouched by this pass — the old sharpest objection stands): one observed instance of lateral spread, or an honest failure to find one after really looking. The live search found zero instances.
3. The PPA re-pricing signature: the market exists at £150–£300/day; whether an enrichment-cover invoice can lawfully and practically deliver mandated coverage, and whether a head will sign that sentence, remains unasked.
4. Concessions printed (unchanged): inherited by rung 0 as spec.

(The no-attainment-claims discipline stands, and now has an external audit date: the outdoor-science RCT reports around late 2026.)

The sharpest new objection: results-compatibility bought no protection

The trained index offered Ashley as the comic existence proof two miles from Max's school. The live record makes it the domain's memento mori. Eighteen years of nature-principled curriculum, top-16% results, an Outstanding judgment — and the academy trust removed the head anyway, on allegations the regulator dismissed in full, after which the practice vanished so completely that the school's own website now carries no trace of it. The walls ledger filed MAT standardisation as slow enclosure, "decades-grained"; the observed instance fired in a single autumn term, against exactly this practice, in exactly this town, and the results record was no defence. What this does to the design cuts both ways and both are printed: the stowaway ladder is a smaller target than a hero-head's whole-school identity — one teacher's afternoon practice never becomes something the trust must have an opinion about, which is now an argument for the low-profile ratchet — and that same argument is unfalsifiable comfort until rung 4 is observed, because the ladder's persistence assumption (one teacher accumulating normalisation for years) lives inside the same institutional weather that ended Harmony at Ashley.

Not yet promoted into propositions.ts — capture, not climb. Forming register; scaffold name. Climbs when: (1) condition zero discharged — the evidence legs re-run live discharged 2026-07-03 (the evidence pass above; the Ashley record verified from the desk, the walk itself still worth taking for the on-foot residue — oversubscription, green-space survey); (2) the wellies arithmetic measured in at least one real classroom; (3) one observed instance of rung 4 — a second teacher adopting because the first one's planning folder got thinner — or an honest failure to find one after really looking, either of which teaches more than this run could.