docs(prd): sandbox-vs-prod posture + certified-accounting-grade operations

New compliance.md leaf: French bookkeeping obligations mapped to fleet
mechanisms — inaltérabilité (L.123-22) via an append-only production
ledger grammar (create/validate/pay/avoir, never mutate a validated
document) enforced by a promote-plan compliance linter; FEC (L.47 A
LPF) with quarterly export + Test Compta Demat validation (accounting-
module binding flagged as unverified gap); piste d'audit fiable (289
VII CGI) framed as a by-product of journals + tx-id reco + monthly
packs; retention, numbering, copie fiable; loi anti-fraude scoped out
(B2B-only) with BlockedLog as sandbox-first belt-and-braces.

New Environments section in agent-architecture: prod = the ledger
(grammar-bound), sandbox = disposable iso-prod rehearsal (exempt, never
wired to production third parties); side_effect_class -> environment/
credential mapping; POCs write on sandbox only; evals target fresh
checkpoints; irreversible-by-design features trial on checkpoints.

Woven through hub (goal, requirement, success criteria, leaves table),
T03/T05/T15 guardrails, QA (linter suite, pure-append snapshots, FEC
cadence, PAF evidence framing), C2, POC-1 exit criteria.

Co-Authored-By: Claude Fable 5 <[email protected]>

vibe/PRD/ai-back-office/README.md

@@ -33,6 +33,7 @@ A **single operator wearing three hats**, plus the fleet itself:
 - **Atomic excellence**: each capability is one narrow, contract-bound atom (extract, validate, record, reconcile, report) that does its one job measurably well. Formats are guaranteed by **deterministic validators, not by model goodwill** — the LLM proposes, code disposes.
 - **The right model for each job** across four tiers — Claude (frontier reasoning), Mistral (EU cloud), local model on the M4 MacBook, SLM on the Raspberry Pi cluster — with graceful degradation when a tier is unavailable. See [model fleet](model-fleet.md).
 - **Human-gated writes as an invariant**: every ERP mutation is rehearsed on the sandbox and promoted through the existing ADR-0003 gate; approvals and digests flow through Telegram. See [agent architecture](agent-architecture.md).
+- **Ledger-grade compliance**: production is operated to the discipline expected of certified French accounting software — validated documents are immutable, corrections are new documents (avoirs), the FEC is producible on demand, and the piste d'audit fiable falls out of the architecture. The sandbox stays exempt *because* it is disposable. See [compliance](compliance.md).
 - **Efficiency**: routine admin costs the human ≤ 15 minutes/day (review + approvals), with hard deadlines never carried in a human head.
 - **Resilience**: no single point of failure — a cloud outage degrades to local triage + queueing, every write is replayable from manifests, books are restorable (tested backups) and provable (content-hashed snapshots).
 - **Institutional memory**: what the fleet learns, decides and audits is distilled into the operator's **second brain** (the PARA Obsidian vault, already live and agent-automated) following its existing conventions — knowledge compounds instead of evaporating into chat logs. See [T17](task-inventory.md#t17--knowledge-capture--retrieval-second-brain).
@@ -114,6 +115,7 @@ flowchart TB
 - **[Agent architecture](agent-architecture.md)** — atom contracts, pipeline shape, write safety, security model (least-privilege ephemeral ERP credentials), prompt-injection defenses, runtimes/scheduling, and the human channel.
 - **[Model fleet](model-fleet.md)** — the four tiers, routing policy, structured-output enforcement, availability model, degraded modes, and cost envelope.
 - **[Challenges](challenges.md)** — the twelve identified risks and their mitigation strategies (the technical "second temps" of this PRD).
+- **[Compliance](compliance.md)** — the French bookkeeping obligations (inaltérabilité, FEC, piste d'audit fiable, numbering, retention) mapped to fleet mechanisms; the production ledger grammar and its linter; the sandbox-vs-production operating posture.
 - **[POC plan](poc-plan.md)** — feasibility proofs as real implementations, ordered, with exit criteria.
 - **[QA strategy](qa-strategy.md)** — golden sets, eval harness, autonomy promotion gates, parity checks, and ops QA. Mandatory per PRD convention.
 
@@ -133,6 +135,7 @@ flowchart TB
 - **Bank**: weekly reconciliation with zero unexplained deltas older than 7 days.
 - **TVA**: every declaration prepared ≥ 5 days before its deadline; dry-run figures match filed figures exactly (€-parity).
 - **Write safety**: zero prod writes outside the manifest → gate → promote path; 100 % of writes replayable from journals.
+- **Ledger discipline**: zero mutations of validated documents (snapshot-verified — corrections exist only as avoirs); the FEC exports clean quarterly once the accounting-module binding is verified.
 - **Resilience**: triage and reminders keep running through a full cloud outage (Pi tier alone); monthly restore drill passes.
 - **Cost**: cloud inference spend ≤ 30 €/month at current volumes (alert at 20 €).
 
@@ -162,6 +165,7 @@ Golden datasets built from real history (mails, invoices, filed declarations), a
 | [Agent architecture](agent-architecture.md) | Atom contracts, pipeline shape, write safety, security, injection defenses, runtimes, human channel. | 🟡 In design |
 | [Model fleet](model-fleet.md) | Four tiers, routing policy, structured outputs, availability, degraded modes, cost. | 🟡 In design |
 | [Challenges](challenges.md) | Twelve risks with mitigation strategies and residual ownership. | 🟡 In design |
+| [Compliance](compliance.md) | Bookkeeping obligations → mechanisms; ledger grammar + linter; sandbox-vs-prod posture; Dolibarr verifications. | 🟡 In design |
 | [POC plan](poc-plan.md) | Ordered feasibility proofs with exit criteria and challenge coverage. | 🟡 In design |
 | [QA strategy](qa-strategy.md) | Golden sets, eval harness, promotion gates, parity checks, ops QA. | 🟡 In design |
 | [STATUS](STATUS.md) | Foundation ledger (shipped PRs) + phase tracker. | 🟢 Current |