SOP Register

The app's standing procedures, mapped to the world's governing bodies · 2026-08-23 edition · Mirrors SOP_REGISTER.md @ 285ab26 — update both.

Why this document exists. The whole app was built on one idea: an app is not a bag of features — it is a set of Standard Operating Procedures, executed faithfully, gate by gate, ending in the professional's real deliverable. If that is true, the SOPs must be listed, so any user, auditor, or regulator can read what the app binds itself to and compare it with the actual SOP of the governing body in their own country. Every SOP below is backed by stage gates and tests in code, not aspiration; its machine-readable form is the Sūtra (versioned, amendable, certified per run).

Part A — Core SOPs (every workflow, every persona)

#	SOP	What it binds	Backed by
A1	Evidence-only answering	No claim without cited evidence IDs; validated against the retrieval set	Claim–evidence contract, evidence gate (ship/downgrade/refuse/surface-conflict)
A2	Conflict preservation	Conflicting evidence shown with both sources — never averaged	Contradiction register + the "never average a conflict away" doctrine
A3	Reserved human decisions	Merges, root causes, dispositions, approvals are made by a person, never auto-asserted	Sūtra humanDecisions + conformance check
A4	Standard of proof before findings	Findings cannot be approved without a declared standard	EvidentiaryStandard gate in handoff
A5	Source reliability rating	Sources rated A–F × 1–6 before weight is placed on them	AdmiraltyCode (NATO/Admiralty scale — international)
A6	Chain of custody	Custody recorded contemporaneously; evidence hashed early	CustodyEvent per SWGDE/NIST — intl. equivalent ISO/IEC 27037
A7	Privacy by architecture	No document, question, or ledger content leaves the device	PrivacyGate + sandbox; designed to satisfy GDPR/CCPA/DPDP by never collecting data
A8	AI disclosure at point of use	Every answer and report declares AI assistance and requires human review	LegalNotice on answers + reports (EU AI Act Art. 50 posture)
A9	Document history	Every deliverable carries its audit trail, printed on the hardcopy	StudioAudit appendix
A10	Amendment-only change	SOPs change by versioned amendment, never silent rewrite; certificates cite the version verified against	SutraAmendment + conformance citation

Part B — Workflow SOPs by persona, with governing-body mapping

Legend: ✅ named in-app · ≈ principle matches, local name differs · ⬜ not yet mapped

Workflow (studio)	Standard named in-app	US	UK / Commonwealth	EU (civil law)	India	International
Investigator — RCA	8D dual root cause (occurrence + escape)	✅ 8D / CAPA	≈ same (8D is global industry)	≈ same	≈ same	✅ ISO 9001 CAPA family
Investigator — ACH	Heuer, Psychology of Intelligence Analysis	✅ CIA tradecraft	≈ UK PHIA structured analytic techniques	≈	≈	✅ NATO Admiralty scale for sources
HR / Compliance	Balance of probabilities; notice + opportunity to respond	✅ EEOC-style practice	≈ ACAS Code (UK)	≈ works-council norms vary	≈ POSH Act inquiry rules	≈ ILO fair-procedure principles
Lawyer — privilege log	FRCP 26(b)(5)(A)	✅	≈ CPR PD 57AD disclosure (E&W)	⬜ privilege differs fundamentally (in-house counsel)	≈ BSA 2023 / Evidence Act privilege	—
Forensic accountant	FRCP 26(a)(2)(B), Daubert, named tracing methods	✅	≈ CPR Part 35 expert duties	≈ court-appointed-expert model	≈ BSA expert evidence	≈ IFAC/ISA 620 using expert work
SIU — insurance fraud	NAIC Model #901, NICB indicators, good-faith referral	✅	≈ ABI/IFB (UK)	≈ Insurance Europe practice	≈ IRDAI fraud framework	—
Journalist — fact-check	Verification + right of reply + alleged-labelling	≈ SPJ Code	≈ IPSO Editors' Code (UK)	≈ EU media councils	≈ Press Council norms	✅ the disciplines themselves are the global newsroom canon
Researcher — review	PRISMA 2020 + GRADE	✅	✅	✅	✅	✅ genuinely worldwide (WHO/Cochrane use both)
Genealogist	Genealogical Proof Standard (BCG)	✅	≈ AGRA/ASGRA norms	≈	≈	≈ GPS is the de-facto global reference
Content creator	Claims/rights/disclosure/corrections	✅ FTC endorsement guides	≈ CMA/ASA (UK)	≈ UCPD + AI Act disclosure	≈ ASCI influencer guidelines	—
Individual — binder	Findability + freshness review	≈ estate-practice norms	≈	≈	≈	universal by nature

Part C — World-coverage audit (the honest part)

Strong globally. The core SOPs (Part A) are jurisdiction-independent — evidence, citation, conflicts, human decisions, custody, privacy are the same disciplines everywhere. PRISMA/GRADE, 8D/ISO-CAPA, the Admiralty scale, and the newsroom disciplines are genuinely international. The architecture (on-device, no data collected) satisfies the strictest privacy regime anywhere by construction, so privacy coverage is worldwide today.

US-named, world-compatible. Where a named standard is US-specific (FRCP, NAIC, FTC, BCG), the procedure the studio follows (describe without revealing; separate fact from opinion; disclose material connections; log nil searches) matches the local equivalent listed above — the local citation is what differs. The report headers currently cite the US instrument.

Real gaps (ranked).

EU privilege — the in-house-counsel privilege difference (Akzo Nobel) is substantive, not cosmetic; the privilege-log studio needs an EU mode note.
Local citation switching — a jurisdiction picker that swaps the cited instrument (FRCP ↔ CPR ↔ BSA) on report headers. Structure is ready; this is a catalog + picker, not a rewrite.
Civil-law standards of proof — the standard-of-proof list carries common-law standards; civil-law "intime conviction" / free evaluation should join it.
Non-English deliverables — hardcopies render in English only.

Verdict: the procedures cover the world; the citations cover the US + international bodies. Items 1–3 above are the path from "US-cited, globally correct" to "locally cited everywhere" — each is a versioned amendment, exactly the change process built for this.

Part D — where to read the SOPs in-app

Settings ▸ System ▸ Sūtra (the Constitution inspector): every phase's obligations, reserved human decisions, and prohibited conclusions, per discipline, with version and amendment history. This register is the human narrative; the inspector is the binding, machine-readable form. Each workflow's jurisdiction disclosure in-app points here.