fix(prose,#17636): relay burst w4 c13 -- rlpt_0d_reward + Lean-10-LeanDojo + 02-2-FLUX-1 (3 carnets, md-only)

[jsboige]

Grain: LIGHT/docs — lane myia-po-2025:CoursIA — prev: LIGHT/docs #18516

Summary

Relais shepherd (burst w4 c13) de la mission prose-counts #17636 — 3 carnets (vagues w4-17 RL, w4-5 symbolicai, w4-7 genai), suppression stricte de mesures d'artefact en prose markdown, recette arbitrée ai-01 (c.5860054240), édition markdown uniquement.

Provenance : commits des workers Vibe (checkpoint-commits locaux, livraison par la lane) :

Carnet	Commit worker	Cellules md	Classification
RL/rlpt_0d_reward_trainer_sota	01f52d218	2	mesures d'artefact resorbées, prédicats gardés
SymbolicAI/Lean/Lean-10-LeanDojo	7a1871302	6	idem
GenAI/Image/02-Advanced/02-2-FLUX-1-Advanced-Generation	e44eb397e	1	idem

Triage anti-double-livraison mesuré : sur les 21 shards porteurs, 13 déjà couverts par les relais ouvertes (#18489/#18496/#18500/#18515/#18516 + #18468), 2 en retrait sur des relais plus récentes (04-2-Transcription derrière #18472 qui a déjà resorbé les tailles ~333 KB ; 03-Claude-CLI derrière #18479/w5 qui a déjà resorbé « en 3 fichiers ») — seuls les 3 carnets non livrés partent ici.

Validation

• Batterie byte-surgical sur le résultat cherry-piqué vs origin/main : 0 cellule code touchée (sources, outputs, execution_count, ids byte-identiques), 9 cellules markdown modifiées, positions inchangées, metadata/nbformat identiques — 3/3 PASS.
• check_prose_quantitative_claims.py --diff origin/main...HEAD : rc=0 (« aucun compteur quantitatif en prose »).
• Aucune PR ouverte ne touche ces 3 chemins (collision guard L898).
• Éditeur md-only : aucune re-exécution due (C.2 non déclenché — aucune cellule source modifiée).

See #17636 (mission en cours, tranche non close par cette PR).

Note G-VAR : genre META (docs) — ne tient pas le plancher G-VAR-1 ; mission coordinator GO ai-01 27/09 (file Vibe), exécutée en file du cycle.

🤖 Generated with Claude Code

[github-actions]

Notebook outputs-required (H.4 schema): PASS (every code cell carries an outputs: list)

[github-actions]

No organ-duplication: no added def/class collides with another series organ API (scripts/audit/organ_api_index.yaml).

Detector: python scripts/audit/detect_organ_duplication.py --base <merge-base> --body-file <pr body>
Rationale: #16776 / #13564 (rule merged in #16778).

[github-actions]

Golden-Set Execution (H.7 P3)

✅ 8/8 notebooks passed (certified reproducible)

Notebook	Status	Time
2.1-Workflow-ML.ipynb	✅ SUCCESS	3.6s
2.2-Descente-de-gradient.ipynb	✅ SUCCESS	4.0s
2.3-Regression-lineaire-logistique.ipynb	✅ SUCCESS	4.6s
2.4-Arbres-Forets-Ensembles.ipynb	✅ SUCCESS	4.5s
Search-01-StateSpace.ipynb	✅ SUCCESS	3.8s
SL-1-LogicalLearning.ipynb	✅ SUCCESS	2.6s
rl_4_multi_armed_bandits.ipynb	✅ SUCCESS	19.6s
GameTheory-04c-NashExistence-Python.ipynb	✅ SUCCESS	3.1s

Pinned lockfile: scripts/notebook_tools/golden_set.lock.txt (H.7 P3, axe A #4208)

[github-actions]

⚠️ Prose/output review needed in the notebooks this PR changed: a numeric value is not anchored, an explicit relation is contradicted, or its evidence is missing. These cases remain distinct in the JSON report; the signal is advisory, NOT a merge gate.

Scope = notebooks CHANGED in this PR, not the whole corpus. Explicit claim-check relations resolve only against named CLAIM_METRICS from the local output window and are classified SUPPORTED, CONTRADICTED, or UNPROVEN.
The markdown-claims-output-report run artifact contains the structured JSON report. See python scripts/check_markdown_claims_output.py --help for re-running locally.
Detector rationale: c.290 / c.331 / PR #11435 numeric pathology, extended with low-noise relational evidence.

[github-actions]

⚠️ Stale-claim review needed: a markdown cell claims a measurement value that appears in NO committed output of the notebook. Advisory, NOT a merge gate — triage against the JSON artifact.

Scope = notebooks CHANGED in this PR, not the whole corpus. The stale-claim-report run artifact holds the structured JSON.
Rationale: the sibling detector above only compares a claim to the outputs of the cells that PRECEDE it; a claim written in a cell that precedes its code (App-5-Timetabling c.2/c.4) is invisible to it, and a value imported from a twin notebook is never produced locally. See python scripts/check_stale_claims.py --help.

[github-actions]

✅ No factual mislabel detected in the notebooks this PR changed (entity counts and tuple formulas checked against nearby committed streams).

Scope = notebooks CHANGED in this PR, not the whole corpus. The factual-mislabel-report run artifact holds the structured JSON.
Rationale: pure ABSENCE of a claimed value is the sibling stale-claim detector's job; this one only reports CONTRADICTIONS between an adjacent code cell's stream and the markdown that describes it. See python scripts/check_factual_mislabel.py --help.

[github-actions]

Trivial-diff advisory (#15740, non bloquant).
genre docs dans la famille META (docs/guard/ledger/readme/test) + diff de 38 lignes changees (<= 100) + aucune exception ecrite dans le body : le litmus de la trivialite (une douzaine d'instances scannees a la suite) est credible. Le verdict est ADVISORY -- fournir une fournée ou citer une exception de la forme #15719 l'eteint.
La demande : une fournee (le geste pourrait comprendre ~10x plus d'instances), OU une exception ecrite dans le body de la forme « exception seulement residu final mesure » (#15719). Editer le body re-deroule cet organe et retire le label.

[github-actions]

Notebook PR Validation: PASS

• Notebooks checked: 3
• Code cells validated: 57
• Result: All passed

Checks: H.1 (no errors), H.3 (execution_count), C.1 (no banned patterns)
Non-Python kernels (.NET/Lean): C.1 + errors only (execution_count advisory)
QuantConnect notebooks: C.1 + errors only (require QC Cloud for execution)

[github-actions]

Path-collision (organ #13359/#13615)

Cette PR #18517 (fix(prose,#17636): relay burst w4 c13 -- rlpt_0d_reward + Lean-10-LeanDojo + 02-2-FLUX-1 (3 carnets, md-only)) touche au moins un chemin de fichier aussi modifie par d'autres PRs ouvertes. Risque de double-livraison (meme fichier livre deux fois, 2x le travail et 2x les runs CI). Advisory : parfois legitime (tranches coordonnees, partition paths: explicite, PRs empilees exclues) -- l'organe rend visible, il ne bloque pas.

• faible -- fix(lean,#18440): align print references with #18199 Lean rename (Lean-10 seul) #18440 partage : MyIA.AI.Notebooks/SymbolicAI/Lean/Lean-10-LeanDojo.ipynb

[jsboige]

[ADJOINT PREFLIGHT]
schema: 1
lane: myia-po-2026:CoursIA-3
pr: 18517
head: a1cde06
complete: true
body: read
comments-reviewed: 9
reviews-reviewed: 0
threads-reviewed: 0
threads-unresolved: 0
surfaces-sha256: 7efd8eac34451931d50f17e68f4eafa7d17b8a3132dfbebe4f92eb67dd4c4c03
diff-files: 3
diff-additions: 19
diff-deletions: 19
checks: latest-wins-green
b0: clear
scope: pass
domain: pass
verdict: READY
[/ADJOINT PREFLIGHT]

[myia-ai-01]

[G-VAR-3 OVERRIDE] lane myia-po-2025:CoursIA -- next: notebook-python

Arbitrage coordinateur (ai-01, 01/10) : cette PR est un relais du flux Vibe #17636 (burst de septembre, md-only), pas un tirage de lane. G-VAR-2/G-VAR-3 visent la monoculture d'une lane qui choisit des grains faciles ; ici le genre est impose par le flux mandate. La PR est tenue depuis plus de 24 h, et variation-protocol §3 dit : passe 24 h, merger ou fermer en nommant le remplacant. Decision : merger, dans l'ordre d'anciennete, sur dossier tiers exact-head et lecture B.0. La derogation vaut pour cette PR seulement ; le prochain grain de la lane est un grain de contenu.