chore(memory): BDR-116 amendment (classifier) + EVAL-045 + journal/TODO/contract w4b — feat model-router wave 4-B

.claude/memory/decisions.md

@@ -1414,4 +1414,8 @@ Branch feature/user-writing-web-rules, UNMERGED (human gate).
 - **Amendment (2026-10-11, wave 3-B, commit 604a6c4)**: clause (2) superseded: the dialog carries CONTEXT (skill description = first sentence of its SKILL.md frontmatter, five YAML forms, hardened read: name allowlist, stat kind file, size cap, `clean()` Cc/Cf + code points; agent description from `agent.offer`, bounded cache; phase `about` line, new 120-char field on the 11 routing.json phases, kept in memory not on Route) and the REAL next model id + effort. Options Later / Keep / Change (T3 Later / Keep). Change = model (4 fixed alias labels with tier role) → effort among those the phases headed by that alias offer (ascending, ≤ 4, one → skipped, zero → Later) → scope; the pair maps to an existing PHASE (rows stay phase names; row's current phase wins a tie; same phase = Keep); a pair no phase offers = a new phase by hand. Other anywhere = Later + toast (no echo). A main-row change toasts the real decision + `/route switch on` hint on a held downgrade. Three lenses rejected inline-route rows and dialog phase edits (2 BLOCKERs). First real use: user moved security-auditor to `judge` (opus/xhigh) through it; floor aligned (`model: opus`, 2026-10-11). Links [[EVAL-044]].
 - **Amendment (2026-10-11, wave 3-C, commit cd8d72f)**: clause (6) extended: `/route forget <name|all|projects>` (composer only) clears decisions in the tracked file through the same writer: a name in every table (`confirmed`, `changed` → row restored to its recorded `from` when it is a string, the row still equals `to` and `from` is a known phase; `projects[*]` pruned), asked again at once; `all`/`projects` confirm through the engine dialog (Cancel first, anything else = cancelled) and hold the single-dialog slot; nothing written when nothing to forget; the answer names a restored row's file + values to realign (a floor aligned to the decision makes the census FAIL after the restore). No per-user decisions file (user: "seulement le forget"): another user inherits the shared memory and may forget. Residuals (security LOW/MEDIUM): raw error text in the answer, file fragments unsanitized in the answer, `floorFile` name not allowlisted, `__proto__` plan/apply mismatch, `changed.phases` restorable. Links [[EVAL-044]].
 - **Amendment (2026-10-11, wave 4-A, commit 68d7ea6)**: logging rule: `verbose` (boolean) logs TRANSITIONS only, built from the decision itself inside `safely` before `next`: main "fable/high → fable/xhigh (<phase>: <cause>[, <model cause>])" once per turn per transition (causes: skill /<name>, route tool, prompt rule, prompt floor, dispatch, sticky /route, session; model: upgrade, switch, fallback: <alias> down); one hand-off line per agent dispatch deduped per turn per (type, target) "main … → agent <type> <alias>/<effort> (row <phase>|explicit|fallback|frontmatter, row|explicit|engine)"; later agent effort moves; an answered-by anomaly (agents only, canonical ids, once per agent+model). The host prefixes the plugin name: new lines carry none. The four stream lines are gone. Residuals W3-A/B/C closed: failure texts to the log, file fragments cleaned at render, floorFile allowlisted, `__proto__` refused, changed scan = rows, dialog Change offers shipped phases only (`mem.filePhases` = validated seed layer), post-answer write failure keeps the key asked. Rejected: a `trace` level (unrequested surface); detecting agent moves at turn.step (decided at spawn). Links [[EVAL-044]].
-
+- **Amendment (2026-10-11, wave 4-B, commit e54a2b5)**: prompt classifier closes the main-loop gap (plain prompt, no skill, no agent, no route in force). Composer `prompt.submit`, trimmed text ≥ 12 chars not starting with `/`, no prompt rule matched, no sticky/turn/run/floor route → one `$.model.classify(sample, [design, analysis, coding, bookkeeping, chat])`, sample = whole text ≤ 2000 code points else head 500 + tail 1500, no `model` option (engine small model). Label → phase (design→plan, analysis→reflect, coding→implement, bookkeeping→apply, chat/unknown → nothing) written as the TURN route, EFFORT ONLY (no tier: main model never moves on a guess), may lower, replaced by any later route tool call or typed skill, ignored at a step whose engine effort ≠ session baseline (`engineBase`: a user `/effort` beats the guess). Bounded: race with `$.clock.sleep(3000)`; timeout/rejection count, user abort neither counts nor writes, every guard re-checked after the await (a route landing meanwhile wins); 3 failures in a row → pause `cooldownMinutes` (×60000, kept across `/clear`, cleared by `/route reload`, one toast + status marker ` · classifier: paused`); skipped while the haiku id is marked down or main runs on haiku. No dialog, no routing.json write on a classified route. Config key `classify` (default true, `pickSwitch`), `/route classify on|off` session toggle surviving a rebuild, `/route show` `classify:` field, status `route: classifier <phase>`, verbose line `(<phase>: classifier)`.
+- **Why**: user 2026-10-11: routing must also cover the main conversation when no skill or agent is invoked. Effort-only because a model flip on a 3-s guess would fight the breaker and the user's `/model`; turn-scoped because a guess must never outlive its prompt.
+- **Alternatives rejected**: first-use dialog on classified phases (the TODO line said so; a dialog per plain prompt is noise and the phase is already confirmed through T3); writing the guess into routing.json; classifying mid-turn or sub-agent prompts; unbounded await (robustness BLOCKER r1: the classify call sat on the critical path with no clock race); `classify` as a tracked-file writer (session toggle + config key instead).
+- **Gates**: plan r1 → r3 (simplicity CONCERNS(4), robustness FATAL(6) BLOCKER unbounded await, correctness CONCERNS(4); confirmation robustness CONCERNS(4)); feater DONE (423) + 1 round (trim, argumentHint, repeated undefined: 428, each mutation-proven); GATE 0 MET; verifier ECARTS(2) coverage → CONFORME on #2; security PASS (0 findings; LOW accepted: pasted text can steer one turn's effort up, README says so). Abort paths judged by reading (`[gated]` clarification). Residual: `smallModelBusy` checks `cfg.models.haiku` while classify uses the engine's default small model (mismatch only when the two differ). Live AC9 open.
+- **Refs**: contract `.claude/tasks/contracts/2026-10-11-model-router-w4b-classifier-1858.md`, plan r3 same slug. Links [[EVAL-045]], [[LRN-210]].

.claude/memory/evals.md

@@ -65,6 +65,7 @@ rules:
 | EVAL-042 | 2026-10-10 | model-router W2 plan r1 → r4: 3 lenses (1 BLOCKER), 2 confirmations (1 BLOCKER then 0); W2-A verifier 3× ECARTS on coverage clauses only, W2-B ECARTS(7) → CONFORME; gate A→B read from engine records | second confirmation paid again (BLOCKER on my own r2 slot); compound coverage criterion = endless ECARTS; engine jsonl replaces the live log |
 | EVAL-043 | 2026-10-11 | model-router W3-A: 3 lenses + 1 confirmation (census BLOCKER, project-tree layer dropped), feater DONE + 4 rounds, verifier 3× then re-verify after a REAL security BLOCK (credential fragment in the tracked key) | challenge + security gates both earned their cost; coverage-shaped criteria still cost 3 verifier rounds; live mod side effects misread as a test leak |
 | EVAL-044 | 2026-10-11 | model-router W3-B (dialog with context): 3 lenses converged on 2 BLOCKERs of my r1 (inline rows, phase edits) + 1 confirmation; feater DONE + 2 coverage rounds; verifier 3× to CONFORME; security PASS; user's first real decision through the new dialog | the lenses pay most when the plan adds a data shape; coverage criteria still cost rounds; live dialogs during the gates must be announced and their data reconciled before commit |
+| EVAL-045 | 2026-10-11 | model-router W4-B (prompt classifier): 3 lenses (BLOCKER: unbounded classify await) + 1 confirmation; feater DONE + 1 round; verifier ECARTS(2) → CONFORME; security PASS 0 findings; LRN-210 applied (one coverage clause per criterion) | first wave since W2 with no verifier cap: per-clause criteria + abort paths declared unverifiable up front cut the rounds to one |
 
 ---
 
@@ -411,3 +412,9 @@ Dogfood: 3 blind lenses attacked the v1 plan for the plan-challenge feature itse
 - **Anomaly**: my r1 proposed a second row shape to satisfy "model then effort" literally; the derived-effort list satisfies it with zero new data shapes. During the gates the user answered the NEW dialog live: security-auditor → opus/xhigh (the request's example), plan-challenger → verify (reset on user go); my reset of `changed` wiped a legitimate entry and left the committed file census-red for one commit.
 - **Action**: when a request says "choose X then Y", look for the existing shape that already encodes (X, Y) before adding one; reconcile routing.json decisions one by one (never `changed = {}`), and run the census before every data commit. Links [[EVAL-043]], [[BDR-116]], [[LRN-212]].
 
+## EVAL-045 — model-router W4-B: per-clause criteria and a declared-unverifiable path ended the verifier loop in one round
+- **Date**: 2026-10-11
+- **Output checked**: plan `.claude/tasks/plans/2026-10-11-model-router-w4b-classifier-1858.md` r1 → r3; diff e54a2b5 (register.ts +211, 428 kit tests, 4 docs).
+- **Method**: simplicity CONCERNS(4), robustness FATAL(6) (BLOCKER: the classify await sat unbounded on the prompt's critical path → `$.clock.sleep` race, abort not counted, post-await re-check), correctness CONCERNS(4) → r2; confirmation robustness CONCERNS(4) → r3. Feater DONE (423, 5 mutation proofs); GATE 0 MET; verifier #1 ECARTS(2) (no test on `trim`, none on the argumentHint) → feater (+5 tests, each failing under its mutation; first trim tests passed without the trim because the length guard caught them, lengthened) → verifier #2 CONFORME 8/8, criterion 9 gated. Security PASS, 0 findings, 1 LOW accepted.
+- **Anomaly**: none on the loop. Verifier observations worth keeping: the sub-agent test proves only spawn+step never calls classify (`prompt.submit` has no agent form in the API types); the 3-s tests pin the upper bound only; the post-await re-check is tested through the route-landing case alone. Security note: `smallModelBusy` guards `cfg.models.haiku`, classify uses the engine small model.
+- **Action**: keep the W4-B contract shape as the template (one clause per criterion, kit-untestable paths listed in CLARIFICATIONS before the first verifier); the live AC9 check is the real oracle for label quality (drop `bookkeeping` if it misfires). Links [[EVAL-044]], [[BDR-116]], [[LRN-210]], [[LRN-212]].

.claude/memory/journal.md

@@ -600,3 +600,4 @@ rules:
 - model-router W3-B (dialog with context) landed 604a6c4 + data 32b71d2 + floor chore. User asked after W3-A: explain what is decided, OK/change/later, model then effort then scope. Plan r1→r3: all three lenses killed my inline-row format and the T3 phase edit (2 BLOCKERs avoided); efforts derived from the phases of the chosen model. Feater DONE first pass (229 tests), verifier ECARTS(2)→(1)→CONFORME, security PASS. GATE 0 heredoc CHECK not runnable again → script. Live during the gates: user answered the NEW dialog for security-auditor → opus/xhigh (the request's example; kept, floor aligned to opus, model-routing lock updated), plan-challenger → verify (reset on user go), orchestrate Keep (T3 check done). My `changed = {}` reset wiped a legit entry → census red for one commit → restored. Docs P1-P6 approved (patch in flight), BDR-116 amendment + EVAL-044 written. Next: doc commit, memory commit, user merge of feature/model-router-confirm (W3-A + W3-B) + publish by hand; T1 live check still open.
 - model-router W3-C `/route forget` landed cd8d72f (+ data commit: escalate Keep). User asked whether decisions can be cleared and what another user can do; chose the command only, no per-user file. Plan r1→r3: three lenses CONCERNS (false `st.mem` premise, kind order, docs scope, census red after restoring an aligned floor, no confirmation on `all`), confirmation CONCERNS(3). Feater DONE; verifier ECARTS(2) with a REAL typo (`ForgetPlan`), then coverage only ×2 → cap → user accepted; security PASS (1 MEDIUM + 4 LOW parked). My own `route(escalate)` call opened the T3 dialog the user Kept. Branch feature/model-router-confirm now carries W3-A/B/C (14 commits), unmerged, unpushed; merge on signal; full make test running.
 - model-router W4-A landed 68d7ea6 on feature/model-router-w4 (+ data commit). User: fewer logs (one line per real change with a cause), residuals, then the classifier. Plan r1→r3: three lenses (BLOCKER: an agent's model move is decided at spawn; `trace` level rejected as unrequested surface; reuse lastPlan/Loop; cause from the decision), confirmation CONCERNS(5) (hand-off line per dispatch deduped per type+target). Feater DONE (341 tests) + 4 rounds; verifier ECARTS(3) real (dedupe on text, double host prefix) then coverage ×2 → cap → user accepted; security PASS + 2 MEDIUM (raw skill names) fixed. Live: the terminal logs changed as soon as the feater saved; user Kept plan-challenger (judge) and the judge phase. Next: W4-B classifier; push-guard hotfix queued.
+- model-router W4-B prompt classifier landed e54a2b5 on feature/model-router-w4. Plain composer prompt with no route in force → `$.model.classify` (5 labels, head 500 + tail 1500) → turn effort only, 3-s clock race, 3 failures pause, `/route classify on|off`. Plan r1→r3 (robustness BLOCKER: unbounded await), feater DONE + 1 round (trim, argumentHint), verifier ECARTS(2) → CONFORME, security PASS 0 findings. No live routing.json change this wave. Branch carries W4-A + W4-B (5 commits), unmerged, unpushed: merge on signal, user publishes. Live AC9 (plain prompts labelled) open.

.claude/tasks/TODO.md

@@ -34,13 +34,14 @@ Plan `.claude/tasks/plans/2026-10-10-model-router-w3a-confirm-1201.md` r4, contr
 - [x] W4-A transition logs + residuals (2026-10-11, 68d7ea6, contract `2026-10-11-model-router-w4a-logs-residuals-1637`, plan r3): verbose = transitions only (main line, hand-off per dispatch, agent moves, answered-by anomaly), stream lines removed, host prefix respected; residuals of W3-A/B/C closed (a)-(g). 3 lenses (BLOCKER: agent moves decided at spawn) + 1 confirmation, feater + 4 rounds, verifier at the cap on coverage (user accepted), security PASS + 2 MEDIUM fixed. Kit 284 → 358.
 - [ ] W4-A residuals (LOW, accepted): `readLayer` logs `String(err)` of a JSON parse error (a file fragment can appear in the local log); `logCall`/`noteMain` ui calls on the step path sit outside `safely` (fail-open on routing, no security impact).
 - [ ] push-guard hotfix: the PreToolUse hook denies a read-only command whose TEXT contains the push verb next to `git` (grep pattern, heredoc); match the command's own argv only.
-- [ ] W4-B classifier (next): haiku `$.model.classify` on typed prompts → turn default route (plan/reflect/implement/apply/chat), first-use confirmation per phase, `/route classify on|off`, transition cause `classifier`.
+- [x] W4-B classifier (2026-10-11, e54a2b5, contract `2026-10-11-model-router-w4b-classifier-1858`, plan r3): `$.model.classify` on plain composer prompts with no route in force → turn effort only (design→plan, analysis→reflect, coding→implement, bookkeeping→apply, chat→nothing), 3-s race, abort not counted, 3 failures pause `cooldownMinutes`, no dialog, no file write, `classify` config key, `/route classify on|off`, status + verbose line. 3 lenses (BLOCKER: unbounded await) + 1 confirmation, feater + 1 round, verifier CONFORME (2nd), security PASS. Kit 358 → 428. Deviation from the TODO line: no first-use confirmation on classified phases (noise per prompt; T3 already confirms the phase).
+- [ ] W4-B residuals: `smallModelBusy` checks `cfg.models.haiku` but classify uses the engine's default small model (guard misses when they differ); abort paths judged by reading only; the 3-s tests pin the upper bound only. Live AC9 open: type a few plain prompts, read `route: classifier <phase>` (verbose line with `/route verbose on`), drop `bookkeeping` if it misfires.
 - [ ] W3-A live checks still open: T1 (typed `/feat` etc.) dialog (T3 seen live 2026-10-11: orchestrate Keep); a parallel same-agent dispatch answered after > 10 s; what an unanswered dialog resolves to on the terminal (must be Later or nothing written). Record in the contract as `[gated]` when seen.
 - [ ] W3-A residuals (security LOW, accepted): non-atomic cross-process write (two sessions, crash mid-write → file reads as failed, previous config kept); `host/path` of a private remote in the tracked file; a persistent write failure re-asks every use (`asked.delete` in the catch); `out.length` vs byte size at the cap; `changePatch` scope 'project' with a vanished key writes Everywhere (cwd change mid-dialog). Deferred: `/route set`, `/route confirm`, `(unconfirmed)` in show, dialog edits of phases, frontmatter auto-alignment, session-start warning when routing.json is dirty, per-machine `projects`.
 - [ ] routing.json ships `confirmed` for verifier/feater (user Keeps) and phases verify/implement: every clone inherits them (by design: decisions travel). Revisit at release if unwanted.
 
 - [ ] W2 residuals: probe 4 unobserved (above); the push-guard hook denies a read-only grep (or a heredoc) whose TEXT contains the push verb next to `git` (security-auditor + orchestrator 2026-10-10, false positives, reworded); ~27 loose "sonnet pin"/"opus pin" shorthand sites kept (true by census); README config key list omits `tiers`/`fallback`/`cooldownMinutes`/`mainUpgrade`/`upgradeMaxTokens` (doc audit item 6, pre-existing); MIGRATION.md "Upgrading to 3.0.0" at release time; `skillCalls`/`spawning` counters global; offers keyed by name.
-- [ ] W3 optional: step heuristics, haiku classifier, quota-aware downgrade, A/B
+- [ ] W3 optional: step heuristics, quota-aware downgrade, A/B (haiku classifier: done W4-B)
 
 ## 2026-09-30 — Higgsfield pack: CLI + skills in the install process, off by default (feature/higgsfield-pack)
 Contract `.claude/tasks/contracts/2026-09-30-higgsfield-pack-1412.md`, spec + plan under

.claude/tasks/contracts/2026-10-11-model-router-w4b-classifier-1858.md

@@ -0,0 +1,32 @@
+# CONTRACT — model-router-w4b-classifier
+- date: 2026-10-11 | flow: feat | branch: feature/model-router-w4
+- status: active
+
+## REQUEST (verbatim — IMMUTABLE)
+okm j'aimerais que tu corrige donc les residus techniques, et aussi que tu fasse cette wave optionnelle du classifieur. [earlier, same day: "est-ce quil y a une comande pour clear…"; "tu me dit le routing de tout skills et agents, mais c'est valable aussi pour la conversatio principale ? Si tu invoaue ni skill ni agebt il faut que ca route aussi on est d'accord ?" → orchestrator: the plain conversation routes only through prompt keywords, ultrathink, derived orchestrate and the route tool; the missing piece is the planned haiku classifier on each typed prompt → user chose it]
+
+## CLARIFICATIONS
+(pass A: none — the W1 plan § W3.2 and the 2026-10-11 exchange define outcome and scope)
+Q: public shape (after three lenses, orchestrator default, user may veto): labels `design | analysis | coding | bookkeeping | chat` mapped to plan / reflect / implement / apply / none; the classified phase becomes the TURN DEFAULT route carrying its EFFORT ONLY (never a model move), and only when NO route is in force (no typed skill, no run slot, no route tool, no sticky `/route`, no floor, no keyword rule, not mid-turn, composer only); no first-use dialog on it (visibility: status line `route: classifier <phase>` and the transition line with `/route verbose on`); `classify` is a machine config key (default `true`, `"classify": false` in `~/.claude/model-router.json` turns it off) with a session toggle `/route classify on|off`; the call is bounded to 3 s and pauses for `cooldownMinutes` after 3 failures; one small-model call per typed prompt of ≥ 12 characters [stated 2026-10-11]
+
+Q: the abort paths (`next.signal.aborted` first in the catch and after the await; classify's own `aborted` cause not counted) cannot be driven by the kit (no abortable dispatch, a throwing bottom hook becomes "no implementation") / A: judged by reading; the live check covers the rest [gated 2026-10-11]
+
+## ACCEPTANCE CRITERIA
+1. On a composer `prompt.submit` with `classify` on, not mid-turn, a trimmed text ≥ 12 chars not starting with `/`, no prompt rule matched and NO route in force on main (no sticky, turn, run or floor), the mod calls `$.model.classify(<first 500 + last 1500 chars>, [design, analysis, coding, bookkeeping, chat])` once, with no `model` option, and writes the turn route `{ effort }` of the mapped phase with source `classifier` (`chat` or an unknown label → nothing); it never runs for a sub-agent prompt, a non-composer origin (so never under `-p`), mid-turn, with the mod off, with `classify` off, when a rule matched, or when a route is in force. One kit test per clause (a bottom `model.classify` mock).
+2. The classified route is turn-scoped (gone at `turn.complete`), carries the effort only (the main model never moves because of it: no tier), may lower the effort (coding → medium, bookkeeping → low), is replaced by any later route tool call or typed skill in the turn, and is ignored at a step whose engine effort differs from the session's baseline (a user `/effort` beats it). One kit test per clause.
+3. Bounded and fail-safe: the call races a 3 s `$.clock.sleep` (timeout → no route, counts as a failure); a user interrupt (`next.signal.aborted`) neither counts nor writes; every guard is re-checked after the await (a route landing meanwhile wins); a rejection or timeout counts, three in a row pause the classifier for `cooldownMinutes` (×60000, kept across `/clear`, cleared by `/route reload`) with one toast and a status marker, a settled answer resets the counter; an unknown label neither warns nor counts; the classifier is skipped while the haiku id is marked down or the main loop runs on haiku. The prompt always proceeds. One kit test per clause.
+4. No dialog and no file write: a classified route opens no first-use dialog and never writes routing.json. One kit test (no `ui.ask`, no `fs.write` on a classified prompt).
+5. Config and command: `classify` is a Config key (default `true`) read from the machine override through `pickSwitch` (invalid → default + one log); `/route classify on|off` is a session toggle (composer only) that survives a rebuild like `switch`/`verbose`; `/route show` prints `classify: on|off` on the `router:` line; the argumentHint lists it. One kit test per clause.
+6. Visibility: the status line reads `route: classifier <phase>` and the main transition line "(<phase>: classifier)" (with verbose on). One kit test per clause.
+7. Kit suite, mods suite, validate and census green.
+   CHECK: cd mods/model-router && out="$(claude plugin test . 2>&1)" && printf '%s\n' "$out" | grep -qE '[0-9]+ pass' && ! printf '%s\n' "$out" | grep -qE '[1-9][0-9]* fail' && claude plugin validate . 2>&1 | grep -q 'passed' && cd ../.. && make test suite=lib/tests/mods.test.sh 2>&1 | grep -q 'all suites green' && bash lib/tests/effort-routing.test.sh >/dev/null 2>&1 && echo W4B-GREEN
+   EXPECT: W4B-GREEN
+   EVIDENCE: MET exit=0 marker-found :: W4B-GREEN
+8. Docs: README (`classifier` paragraph in the model-router section: what, cost, `/route classify`), USAGE (`/route` row + one sentence in "Niveau d'effort"), CHANGELOG Unreleased Added, `lib/effort-shift.md` (≤ 2 lines).
+   CHECK: grep -qi 'classif' README.md && grep -qi 'classif' USAGE.md && grep -qi 'classif' CHANGELOG.md && grep -qi 'classif' lib/effort-shift.md && echo W4B-DOCS
+   EXPECT: W4B-DOCS
+   EVIDENCE: MET exit=0 marker-found :: W4B-DOCS
+9. Live check (orchestrator + user, after the gates, `[gated]`): a few plain prompts typed in the terminal, each with the label it got (status line `route: classifier <phase>`, transition line with verbose on); drop `bookkeeping` if it misfires.
+
+## FILE SCOPE
+mods/model-router/hooks/register.ts, mods/model-router/hooks/register.test.ts, README.md, USAGE.md, CHANGELOG.md, lib/effort-shift.md, lib/tests/effort-routing.test.sh (only if a lock needs it)

.claude/tasks/plans/2026-10-11-model-router-w4b-classifier-1858.md

@@ -0,0 +1,172 @@
+# PLAN r3 — model-router wave 4-B: prompt classifier (2026-10-11)
+
+r1 → r2 after simplicity CONCERNS(4), robustness FATAL(6) (BLOCKER: the
+classify await is unbounded on the critical path), correctness CONCERNS(4).
+r2 → r3 after the confirmation (robustness CONCERNS(4)). Precedence
+r3 > r2 > r1. Contract `.claude/tasks/contracts/2026-10-11-model-router-
+w4b-classifier-1858.md`. Base: feature/model-router-w4 at 86845fc.
+Corrected facts: default keyword rules never run mid-turn and only write
+plan/reflect (they never lower); `mainEffort` ranks sticky > turn > run >
+floor > engine and `mainModel` ranks turn above floor; `runMain` holds a
+typed best-tier skill's route across the run's turns; `$` calls do not
+count against the hook budget (an unbounded `classify` await freezes the
+prompt), while `$.clock.sleep` does; `ClassifyOptions` has `model` only;
+`claude -p` prompts arrive with origin `sdk`; the breaker is fed by
+`classic.StopFailure` only; `warnOnce` fires once per session; config
+switches (`enabled`, `verbose`, `mainModelSwitch`) are Config keys read
+through `pickSwitch` with session toggles; `confirmPhase` (T3) is
+confirm-only, so a dialog on a classified phase would decide nothing.
+
+## Decisions
+- D1 no first-use dialog on classified phases (Later and Keep would apply
+  the same route). Visibility = the status line (`route: classifier
+  <phase>`) and the W4-A transition line (needs `/route verbose on`).
+- D2 `classify` is a Config key (default `true`), read from the machine
+  override through `pickSwitch` (invalid → default + one log), with a
+  session toggle `/route classify on|off` next to `switch`/`verbose`,
+  printed on the `router:` line of `/route show`. Persistent off =
+  `"classify": false` in `~/.claude/model-router.json`. routing.json is
+  NOT touched; no new writer.
+- D3 the classifier fills only an EMPTY route: it runs only when no route
+  is in force on main (`mainRoute(st) === null && st.turnFloor === null`:
+  no sticky `/route`, no turn route from a rule/skill/tool, no run slot
+  from a typed skill, no floor), evaluated after `routeFromPrompt`. Real
+  coverage, stated in the docs: a typed best-tier skill (feat, bugfix,
+  ship-feature…) holds its run slot across turns, so nothing is
+  classified until `/route clear`, `/route off`, a user `/model` or a
+  typed non-best skill; a non-best skill's row is turn-scoped, so its
+  follow-up prompts ARE classified (accepted: those rows are low/medium
+  and a complex follow-up may rightly rise). Never mid-turn (`e.turnId !==
+  undefined` → skip, as the rules). Never for a non-composer origin (so
+  never under `-p`). Never for a sub-agent prompt (prompt.submit is main
+  only; a test proves an agent's prompt is not classified).
+- D4 a classified route carries the EFFORT ONLY: `turnMain = { phase,
+  route: { effort: cfg.phases[phase].effort }, source: 'classifier' }`.
+  No tier, so the main model never moves because of a guess (no cold
+  context re-reads on opus/sonnet sessions, no cost surprise). It may
+  lower the effort (coding = medium, bookkeeping = low); turn-scoped.
+  An explicit user `/effort` beats it: `st.engineBase` = the engine
+  effort of the first main step of the session (kept across /clear);
+  at `turn.step`, when `e.effort !== st.engineBase` (the user moved the
+  builtin effort) a `classifier`-sourced turn route is ignored for that
+  step (the engine effort applies) — "un choix explicite bat la phase
+  déduite" (BDR-115 amendment b).
+- D5 bounded call: `label = await Promise.race([classify, $.clock.sleep(
+  CLASSIFY_MS = 3000, { signal: next.signal }).then(() => TIMEOUT)])`
+  inside try/catch: in the catch, `next.signal.aborted` is tested FIRST
+  (a user interrupt: no count, no write, return); a TIMEOUT counts as a
+  failure (D7) and gives no route; after the await every guard of D8 is
+  RE-CHECKED (`st.off`, the switch, no route in force, not paused) before
+  the write, and `next.signal.aborted` skips it (a late result never
+  lands in a later turn). The losing classify promise is left to settle
+  (the kit offers no cancel); stated.
+- D6 labels are descriptive and mapped: `design` → plan, `analysis` →
+  reflect, `coding` → implement, `bookkeeping` → apply, `chat` → none
+  (lookup with `hasKey`; a mapped phase without an effort → no route).
+  The call passes no `model` option (the engine's small fast model; no
+  alias from a hook, LRN-203). The sample = the whole trimmed text when
+  ≤ `CLASSIFY_MAX = 2000` code points, else the first `CLASSIFY_HEAD =
+  500` + last `CLASSIFY_TAIL = 1500` code points (cut by code points,
+  never a lone surrogate). A pasted tail can steer the label for one
+  turn (effort only, no file write): accepted and documented.
+- D7 failures: a rejected call (classify's own `aborted` cause excepted)
+  or a TIMEOUT increments `st.classifyFails`; a settled answer resets it;
+  at `CLASSIFY_FAILS = 3` the classifier pauses until `clock.now() +
+  cooldownMinutes * 60000` (`st.classifyUntil`) with one toast
+  "classifier paused <n> min (3 failures)" and the status marker
+  `classifier: paused`; both fields are in `resetSession`'s kept set and
+  cleared by `/route reload` (like the breaker); the expiry check runs
+  BEFORE the route-in-force guard (the marker never stays stale); an
+  unknown label neither warns nor counts; the classifier is skipped while
+  `cfg.models.haiku` (a stated proxy for the engine's small model) is in
+  `st.down`, and when the main loop itself runs on haiku (the effort is
+  stripped there anyway).
+- D8 guards summary (in order): `st.off` → `cfg.classify` (+ toggle) →
+  composer → `e.turnId === undefined` → not a slash → trimmed length ≥ 12
+  → no rule matched this prompt → no route in force → not paused, model
+  not down → classify.
+
+## Steps
+- [ ] S1 config: `classify` key (DEFAULT true, pickSwitch), toggle,
+      `/route classify on|off`, show `router:` line, argumentHint.
+- [ ] S2 `classifyPrompt($, st, e, next)` in the prompt.submit hook after
+      `routeFromPrompt` and the typed marker; `LABELS`/`LABEL_PHASE`
+      tables; the sample helper; the race; the failure counter and pause;
+      the effort-only route write with source `classifier`; `Source`
+      gains `'classifier'`; `slotWord` maps it to `classifier`; the
+      status line already prints `route: <source> <phase>`.
+- [ ] S3 docs: README (classifier paragraph: what, one small-model call
+      per typed prompt when no route is in force, effort only, `/route
+      classify off` or `"classify": false` per machine, where to see it:
+      status line, `/route verbose on`); the `/route` subcommand lists
+      (README ×2, USAGE) gain `classify on|off`; the machine keys lists
+      (README, USAGE) gain `classify`; USAGE "Niveau d'effort" one
+      sentence; CHANGELOG Unreleased Added; `lib/effort-shift.md` ≤ 2
+      lines (the plain conversation routes by content now).
+- [ ] S4 kit tests (a bottom `model.classify` mock; `$.clock` mock for
+      the race), one per clause: called once with the head+tail sample
+      and the five labels, no model option; each label → its phase effort
+      only (no tier on the route) and `chat`/`undefined` → none; never
+      resolving → no route after 3 s, prompt proceeds; rejection ×3 →
+      pause + toast + status marker, then no call during the cooldown,
+      a success resets the counter; `aborted` → no count, no warn; small
+      model down → not called; not called: short text, slash, sdk origin,
+      mid-turn (`turnId`), rule matched (keyword and ultrathink), sticky
+      `/route`, run slot held, floor carried, `classify` off (config,
+      toggle), mod off, an agent's prompt; turn-scoped (gone at
+      turn.complete); effort lowered to low for bookkeeping; the main
+      model unchanged on a sonnet session for `design`; `/route show`
+      prints `classify: on`; status line `route: classifier reflect`;
+      W4-A line "(reflect: classifier)"; invalid config value; 3 timeouts
+      → pause; a dispatch abort mid-race → no count, no write; a `/route
+      <phase>` landing during the await → the classifier result dropped;
+      pause expiry at cooldown-1 ms (still paused) and +1 ms (resumes);
+      `/clear` keeps the pause, `/route reload` clears it; a user
+      `/effort` change (engine effort ≠ baseline) → the classified effort
+      ignored at the step; a 2000-code-point text sent whole, a longer
+      one head+tail with no lone surrogate; `constructor` label → none;
+      a phase without effort → none; main on haiku → not called; a
+      non-best skill's follow-up prompt classified; a plain prompt after
+      a finished best-tier run not classified until `/route clear`.
+- Edge cases: a prompt that is only a URL or code (the classifier
+  decides; no special-casing); every classified prompt waits one small-
+  model round trip (typically under a second, 3 s at most: stated in the
+  README and measured in the AC9 live check); a `[1m]` session (no
+  effect: effort only).
+- Disposition: honors BDR-115 (no alias from a hook, no model write, no
+  file write), BDR-116 (writers unchanged; no new dialog), LRN-203,
+  LRN-210, LRN-212 (live mod: the first plain prompt after the feater
+  saves is classified).
+- Docs note for the user: a misclassification lowers one turn's effort
+  only; `/route <phase>` or `ultrathink` overrides it at once.
+
+## Challenge ledger (r1 → r2)
+- simp 1, corr 3/7, rob 6 (dialog decides nothing / misstates) → D1 dropped.
+- simp 2, corr 11, rob 11 (flag plumbing, new writer) → D2 Config key.
+- simp 3/4, corr 1/2/12, rob 3/4 (mid-turn, run slot, floor) → D3.
+- rob 5 (model flips per turn) → D4 effort only.
+- rob 1 BLOCKER, corr 9 (unbounded await, late write) → D5.
+- simp 5, corr 4 (alias from a hook) → D6 no model option.
+- corr 8, simp 8 (label words) → D6 descriptive labels mapped.
+- rob 7 (head-only sample) → D6 head + tail.
+- rob 2, corr 5 (failure handling, warn spent) → D7.
+- corr 6, rob 9 (`-p`) → D3 never.
+- corr 10 (sub-agent clause) → S4 test.
+- simp 6, rob 10 (visibility) → status line + verbose note.
+- simp 7 (doc lists) → S3 lists named.
+
+## Confirmation ledger (r2 → r3)
+- conf 1 (timeouts uncounted) → timeouts count toward the pause.
+- conf 2 (abort lands in the counted branch) → `next.signal.aborted`
+  tested first in the catch.
+- conf 3 (classifier overrides a user `/effort`) → `engineBase` guard at
+  the step: an explicit builtin effort beats a classified route.
+- conf 4 (run coverage) → D3 states the real coverage; docs name it.
+- conf 5 (guards not re-checked after the await) → re-checked.
+- conf 6 (pause bookkeeping) → ×60000, kept across /clear, cleared on
+  reload, expiry checked first, boundary tests.
+- conf 7 (sample, labels) → whole text ≤ 2000 cp, code-point cuts,
+  hasKey, no-effort phase skipped.
+- conf 8 (down check) → haiku id as a stated proxy; skip on a haiku main.
+- conf 9 (pasted tail) → accepted, documented.
+- conf 10 (latency) → stated in the README and measured live.