Phase A½. The move from feature to product development: for every assessment, answer "how sure are
we that this answer is COMPLETE?" — different from confidence. The product never claims full coverage;
it makes its own knowledge state transparent and auditable. Shows what we do NOT know and why.
- compliance/completeness/: assess_completeness(identified, corpus_status, uncertain, assumptions,
assessed_obligations) -> CompletenessReport. Separates IDENTIFIED from ASSESSED (validated corpus
AND determined applicability) and justifies every gap. Two kinds of open: corpus gap (future_corpus)
and applicability uncertainty (query_required + deciding question, e.g. Data Act / generates_usage_data).
- The metric is COUNTS, never a single percentage: "Identifiziert N · bewertet M · offen K ·
Unsicherheiten U · Begründung ja" + an honest audit statement.
- ADR-007: auditable honesty; phase order A factory -> A½ Completeness -> B new domains; the
transparency selling point. Deterministic, no LLM; corpus status + obligation count injected.
- reference suite: "Regulatory Completeness" section runs an industrial-dishwasher assessment
(assessed CRA/MaschinenVO; open EMV/Environmental=future_corpus, Data Act=query_required) and notes
Environmental flips open->validated automatically once the corpus lands.
11 completeness tests (54 with adjacent modules), mypy --strict clean (15 files), check-loc 0.
Product code with no app caller + ADR/reference = non-runtime -> no deploy (ADR-001). Freeze-safe.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Phase A1. The real knowledge production is not writing — it is TARGETED UPDATING: when 20 documents
arrive, which 5 change our knowledge and which 15 are ignorable? Before the parser, Knowledge Intake
classifies a new document (no content extraction) and intersects its signals with an index of the
existing knowledge to emit a Knowledge Package (an impact analysis).
- compliance/knowledge_intake/: build_knowledge_index(patterns, playbooks, reference_scenarios,
obligation_index) + assess_document_impact(descriptor, index) -> KnowledgePackage. Deterministic,
NO content extraction, NO LLM. Surfaces affected capabilities / playbooks / transition patterns /
reference scenarios / (injected) obligations, whether it is a new domain, and a triage level
(HIGH / LOW / NONE / NEW_DOMAIN) with a recommendation.
- ADR-006: Knowledge Intake = classify + impact before extraction; full factory Intake -> Package ->
Parser -> Draft -> Review -> Published; phase order A1 Intake / A2 Draft / A3 Review.
- reference suite: "Knowledge Intake" section triages 3 example documents (CRA SBOM-FAQ -> high,
14C/2PB/3RTS/2Obl; environmental guidance -> new_domain; marketing blog -> ignorable). Section
lives in _helpers.py to keep generate.py under the 500-LOC budget.
- Honest known refinement surfaced by intake: regulation-ID normalization (CRA vs Cyber Resilience Act).
10 intake tests (60 with the adjacent modules), mypy --strict clean (16 files), check-loc 0.
Product code with no app caller + ADR/reference = non-runtime -> no deploy (ADR-001). Freeze-safe.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
The bottleneck is not content, it is knowledge PRODUCTION. Instead of writing 200 playbooks by
hand, generate drafts deterministically from data the software already owns, then have an expert
review them. Mirrors the legal pipeline (Gesetz -> Parser -> Obligation -> Review) for BreakPilot's
own knowledge: new Capability -> Registry -> Transition Pattern -> Playbook Draft Generator ->
Expert Review -> versioned Playbook.
- compliance/knowledge_production/: generate_playbook_draft(capability, requirement, control_links)
+ drafts_from_pattern(pattern) -> one PlaybookDraft per delta capability. Owned fields (why /
closes_regulations / expected_evidence / typical_controls) are assembled with per-field provenance;
the practitioner know-how (tools / process_steps / how_others) is left as an explicit TODO.
- DraftStatus lifecycle (Freigabestatus): draft_generated -> in_review -> reviewed -> validated ->
proven. Deterministic, NO LLM in the core (any model enrichment stays offline/advisory/propose-only).
- ADR-005: extends "the engine does not change, the corpus grows" with "and the corpus is not written
by hand — it is deterministically prepared, then curated".
- reference suite: "Knowledge Production" section turns the convergence pattern into 12 auto-assembled
drafts (why/closes/evidence filled, tools/steps TODO) -> review 12 drafts, don't write 12 playbooks.
10 tests (50 with playbook/optimization/transition/company), mypy --strict clean, check-loc 0.
Product code with no app caller + ADR/reference = non-runtime -> no deploy (ADR-001). Freeze-safe.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Roadmap item 4. After WHAT applies / WHAT is missing / WHICH first, the GF asks HOW. The
Implementation Playbook renders, for one capability, the full journey — why / which regulations
it closes / tools / process / evidence / controls — and chains the Optimization Roadmap into
per-measure playbooks. Another renderer over the same Capability spine (ADR-003/004), not a new
engine: ~95% of the data already exists, it just needs a different rendering.
- compliance/playbook/: build_playbook() + playbooks_for_plan() (chains optimization -> playbook,
acyclic; reuses leverage for "closes which regulations"). Capabilities without curated content
render as honest status:missing stubs — the content-owed signal.
- knowledge/implementation_playbooks/: curated knowledge layer (Reasoning Knowledge Acquisition),
two deep expert drafts (SBOM, CVD/PSIRT, status draft, expert-draft-not-normative) + README.
The bottleneck is now CONTENT, not software; Playbook (own knowledge) != regulatory domain.
- ADR-004: Implementation Playbooks = renderer + knowledge layer; content is the bottleneck.
- reference suite: "Implementation Playbook" section renders the SBOM journey + Roadmap->Playbook
table (high-leverage caps flagged "fehlt (Inhalt)" — content backlog, highest leverage first).
- refactor: extracted markdown helpers to reference_scenarios/_helpers.py to keep generate.py
under the 500-LOC budget.
9 playbook tests (40 with optimization+transition+company), mypy --strict clean, check-loc 0.
Product code with no app caller + knowledge/ADR/reference = non-runtime -> no deploy (ADR-001).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Roadmap item 5. GAP analysis and measure-prioritisation are the SAME computation: Required −
Known = the Capability Delta. The Capability Delta Engine (RS-005) computes it once; renderers
read that ONE delta. Interview Renderer (missing info → questions) was already built; this adds
the Roadmap/Management Renderer (missing capabilities → measures ranked by regulatory leverage).
- compliance/optimization/: regulatory_leverage() + select_within_budget() (pure leverage math)
+ roadmap_from_delta(assessment, ...) — the keystone binding optimization to the RS-005 delta
(dependency optimization → transition_reasoning, acyclic; the delta engine stays hermetic).
leverage(measure) = number of regulatory requirements it closes at once (e.g. patch management
→ CRA+MaschinenVO+IEC62443+ISO27001 = 4). No new corpus, no new meta-model class (freeze v1.0).
- Welt-1 honesty: percentages are exact count ratios over the IDENTIFIED requirements (the known
delta), never "% gesetzeskonform".
- reference suite: "Regulatory Optimization" section runs the SAME convergence delta → ranked
measures + budget answer + the management sentence "of N identified requirements you close M
with the top-K measures (X%) — highest regulatory leverage".
- ADR-003: Capability Delta Engine — one delta, many renderers; rename Gap → Capability Delta.
13 optimization tests (31 with transition+company), mypy --strict clean, check-loc 0.
Product code with no app caller + ADR/reference = non-runtime → no deploy (ADR-001).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Aligns the spec with RS-005 v0: the Transition Planning Engine owns the INFORMATION
GAPS (TransitionQuestionRequest), not the questions. Chain: Planning Engine ->
TransitionQuestionRequest -> Question Renderer (RS-005.1) -> Interview. RS-005.1
(renderer/templates) deliberately deferred; GeneratedQuestion reframed as the renderer's
output (a swappable policy layer), not part of the engine.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
v1.1: interview questions are GENERATED from the existing (Master) Controls, not
hand-written. Three building blocks: Control->question_intent (corpus/Execution),
~30-40 Master Question Templates (Reasoning), Transition-Prioritization (certs decide
which generated questions can be skipped; 217->19 funnel, reuses Company 2A + cert map).
v1.2: knowledge production. LLMs produce the first expert DRAFT (the prioritization per
transition); BreakPilot reviews + versions + OWNS the canonical library (in Git, not the
AI; model-independent, MDQ-00127 v4). Offline multi-model workflow, NOT runtime
(deterministic-first: LLM offline-propose, never online-mutate). Hard boundary: the
library is an expert DRAFT, not a normative/legal proof -- "cert probably covers X" is
Welt-1 (ClaimCoverage), never "erfuellt" (anti-fake-evidence).
Reframes the 100 seed questions as validation/template-extraction set. Spec only, no
code; non-runtime docs -> no deploy (ADR-001).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Second reasoning mode (extends, does not replace): BreakPilot answers MIGRATION
questions (start state -> target state -> delta), not regulation Q&A. New package
compliance/transition_reasoning/ (spec only). Transition Reasoning is RCI
generalized; reuses Company 2A (have), Master Capability Registry (MCAP) and RCI.
MDQ Registry = 4th identity-machine instance (after Master Controls/Obligations/
Capabilities): every Master Delta Question is a versioned, identifiable knowledge
unit (verifies MCAP, supports obligations, transition patterns, evidence types,
information gain, confidence impact, follow-up). Transition Patterns hold only MDQ
references -> reuse across transitions. Delta interview = information-gain
optimization, not a sequential questionnaire.
ADR-002: transitions are DATA (patterns + capability/MDQ knowledge), never engine
or metamodel extensions. 100 seed questions captured as v1.
Spec only (no code; freeze-respecting: additive package, no new graph/base class/
meta-model class). Non-runtime docs -> no deploy (ADR-001).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
A dev deploy must always have a verifiable runtime effect. Deploy only on
runtime/API/data-model/reasoning/security changes; docs, reference suites, ADRs,
board and ownership texts are merged to origin/main but NOT pushed to dev (no Orca
build). Keeps the CI/CD history meaningful: every build == a runtime change.
Architecture/release decision (not a developer convention) -> own folder
docs-src/architecture/adr/. Non-runtime: this commit triggers no deploy, per its
own policy.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
User-Reframe (die eigentliche Reife): nicht „Session X besitzt Knoten Y", sondern jede Session
besitzt KANTEN. Edge-Ownership-Tabelle: Feature/Cert->Cap = S3 · Cap->Obligation/Procedure/
Control/Evidence = S2 · Citation-Span->Legal-Basis = S1. Kein Owner hält alle ein+ausgehenden
Kanten eines Knotens. `cap.*` = kanonische ID auf obligation_id-Niveau. Capability = EINZIGER
Knoten über 3 Welten (Recht/Produkt/Nachweis) = semantischer Mittelpunkt. Künftiger Vertrag:
Confidence/Disambiguierung bei mehreren Capabilities = Domaene 3, Domaene 2 vertraut geliefertem
cap.X. Domaene 2 ruht stabil bis Wake-up.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
User-Praezisierung: Domaene 2 ruht NICHT unbestimmt. Wake-up-Trigger (EINER reicht):
Feature Graph>=200 Features · Span-Anker verfuegbar · neue Regulierung ingestiert · Runtime
kennt neue Evidence-Typen. Erster Folgeauftrag (gated auf Feature Library v1):
FEATURE COVERAGE REPORT = Wissenslueckenanalyse pro Feature (Feature->cap.*->Obligation->
Procedure->Evidence -> Coverage %; zeigt fehlende Capability/Procedure/Evidence je Feature).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
User-Antwort auf „wie verteilen wir die Arbeit": nach BESITZ der Datenmodelle, NICHT nach
Regulierung. 3 Domaenen (Legal Knowledge / Compliance Execution / Product Knowledge), jede
besitzt EIN Modell (andere read-only). 3 Vertraege: Legal->Compliance citation_span->legal_basis ·
Product->Compliance Feature->Capability (WICHTIGSTE Schnittstelle) · Compliance->Legal
obligation_id->legal_basis. Product Knowledge Graph = naechster Meilenstein (Reasoning-Session
umfokussieren, besitzt schon CanonicalProductRegulatoryProfile+Navigator). NIS2 verschoben.
Offene Fragen: Legal-KG-Owner, IACE-4.-Session, Compliance-2-Branch-Split.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
User-Entscheidung: Metamodell als v1.0 einfrieren (nur META-SEMANTIK: 6 Klassen + Kanten-
Vokabular + Attribute; NICHT Registry/Capabilities/Procedures). Architektur-Freeze in Kraft:
neue Regulierung = DATEN nicht Architektur; 0 neue Objektklassen erwartet; reopen nur bei
nachgewiesenem Scheitern (Hazard/Threat = einzige bekannte künftige Öffnungs-Ursache, nur fuer
FMEA). Reuse-Metrik-KPI definiert (Wissens-Akkumulations-Beweis). Validiert gegen 5
Regulierungsarten (DSGVO/CRA/MaschVO/Data-Act/NIS2). Erster Live-Durchlauf: MaschVO.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
User-Stresstest VOR der naechsten Regulierung: passt MaschVO/Data-Act/AI-Act/NIS2 ins
6-Klassen-Modell (Obligation/Capability/Procedure/Control/Evidence + Guidance) OHNE neue
Objektklasse? Ergebnis 4x NEIN -> Compliance Meta Model steht. 2 Verfeinerungen
(realized_by Capability OPTIONAL; Risiko-Niveau/Frist/Hazard-Schwere/Risiko-Tier = Attribute,
keine Klassen). 1 Watch-Point: Hazard/Threat (erst noetig bei quantitativem FMEA-Risiko als
First-Class-Knoten, nicht fuer Compliance-Abbildung). Kein Code, keine Regulierung ingestiert.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Schema-Papier statt capabilities.json (User-Entscheidung). Befund: die 8 SHARED_CAPABILITY-
Cluster zerfallen in Typ-1 (technische Capabilities: mfa/tls/code_signing/session/anomaly)
und Typ-2 (Sicherheitsziele: attack_surface_min/software_integrity = die #4-Gaps). Empfehlung
Modell C: Capability = EINZIGE neue Klasse; Sicherheitsziele = CORE Legal Obligations
(CORE/DOMAIN existiert bereits). Kanten-Graph (realized_by/specializes/...). guidance_basis
gehört konzeptionell an die Capability. 4 Entscheidungen offen (User). #5b Materialisierung
GEGATED auf Modell-Annahme — keine Daten verschoben.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
9 docs + index in docs-src/architecture/ documenting the deterministic
retrieval engine: retrieval pipeline, authority rerank, source_class,
source_role, control-intent + diversity, assessment, confidence,
explainability + supersede, framework_* layer. Each doc carries the exact
constants, the rationale behind them, code refs, and the failure class
it addresses. Audit/onboarding reference.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Schreibt das Zielmodell fest: Legal Obligation = gemeinsame Sprache zwischen
Legal Knowledge Graph (Chat) und Compliance Execution Graph (Engine).
- Registry-Schema v1 (id/tier/legal_basis/guidance_basis/facets/citation_anchor_ids/
relationships/decision_method)
- Zitierfähigkeit hängt an der OBLIGATION, nicht an Controls (Regulierungsänderung =
Anchor tauschen, Controls unverändert)
- legal_basis (Primärrecht) vs guidance_basis (NIST/OWASP/...) + source_role
(LEGAL_BASIS/GUIDANCE/EVIDENCE/IMPLEMENTATION/OUT_OF_SCOPE)
- HARTE Regel: LEGAL_MINIMUM nur mit Primärrechts-Anker
- Beziehungsgraph: requires/implements/supports/produces_evidence_for/depends_on/derived_from
- Citation-Anchor-Pipeline Document→Obligation (KEIN Re-Ingest zum Control-Neubau)
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Friert das Kriterien-Meta-Modell ein: atomare getypte Kriterien mit drei
Achsen (verification_method, decision_method, compliance_tier), 3-Status-Gating
nur auf LEGAL_MINIMUM (ERFÜLLT/TEILWEISE/FEHLT), 3-Ebenen-Reporting und
Grün/Blau/Rot-Semantik. Control-UUID bleibt stabil (kein physischer Split),
Speicherung in generation_metadata jsonb (keine Schema-Änderung). Validiert am
Pilot (6/6 Disagreements korrigiert, TEILWEISE empirisch bestätigt).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
ai-sdk (legal_rag_client/scroll/types) liest die gepinnten Spec-Felder
article_label/regulation_code/article/paragraph/sub/citation_style/is_recital
mit Fallback auf alt-ingestierte Chunks (regulation_id, section); neuer getBool-Helfer.
Advisor + Drafting-Engine bilden die Quellenzeile primaer aus article_label
("BDSG § 38 Abs. 1"), sonst aus den strukturierten Feldern. 17 Tests gruen, tsc sauber.
Vertrag: docs-src/development/rag_reingest_spec.md (§2/§7). Deploy an den Re-Ingest
gekoppelt — neue Felder sind bis dahin leer (graceful Fallback).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
11 Modul-Eintraege entfernt, deren exakte Route bereits ein immer-sichtbarer
Pipeline-Schritt ist (advisory-board, ai-act, source-policy, loeschfristen,
einwilligungen, cookie-banner, dsr, vendor-compliance, consent-management,
email-templates, training) — Heimat bleibt die Pipeline, kein Feature-Verlust
(keiner dieser Schritte hat visibleWhen). "Datenschutz"-Gruppe zu "Cookie &
Consent" (Consent Dashboard + Cookie Live-Vorschau) verschlankt. Aehnlich
benannte, aber VERSCHIEDENE Seiten bewusst behalten (document-generator≠
catalog-manager, control-library≠coverage, consent≠consent-management,
cookie-banner≠/preview, vendor-compliance≠vendor-assessment).
Vollstaendige Routen-Inventur (Pipeline + Module + aufgeloeste Dups) in
docs-src/development/sdk-navigation-inventory.md — damit kein Feature
unsichtbar verloren geht.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Jede Normreferenz einer Maßnahme wird lizenzklassifiziert (eu_law /
public_domain / open / paid_reference) — paid-reference-Normen werden nur als
Verweis geführt, nie im Text gespeichert (idea/expression). Kuratierte
Maßnahmen tragen Tier 'core', KI-/Fallback-Maßnahmen 'review' (indikativ).
Frontend zeigt Quellen-Badges + "indikativ"-Kennzeichnung. Methodik in
docs-src/development/mapping-methodology.md (Szenario C, Due-Diligence).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Ergänzt nach Rückmeldung der Controls-Session: ID-Stabilität schützt auch deren
atom_classification (~161k) + addressee (control_uuid-gebunden); deren Step-1/2 ist
additiv (tier/source_type/core_count/addressee, bricht Verträge nicht); eine
Wahrheit — Muster-Schicht aus atom_classification speisen, nicht neu ableiten.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Andock-Vertrag für die Maßnahmen-Schicht: stabile Muster-Einheit + feste ID,
control→pattern-Mapping, Framework-Crosswalk pro Muster. Abstimmung mit der
Controls-Session (core/control-pipeline). CRA-Spine/M5xx bleiben unabhängig.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
check-rebuild-needed.sh war seit Mai funktionsfähig nur fuer 3 von 10
Containern. Die anderen 7 Dockerfiles hatten kein ARG/ENV BUILD_SHA und
docker-compose.yml hat fuer KEINEN Service den Wert durchgereicht — daher
defaultete BUILD_SHA ueberall auf "unknown" und die Drift-Check war
zahnlos.
- ARG BUILD_SHA + ENV BUILD_SHA in 8 zusaetzlichen Dockerfiles
(ai-compliance-sdk, developer-portal, document-crawler, dsms-gateway,
compliance-tts-service, docs-src, docs-site, dsms-node)
- docker-compose.yml: BUILD_SHA: \${BUILD_SHA:-unknown} in jedem build:
Block (10 Services)
- .gitea/workflows/ci.yaml: neuer Job build-sha-integrity validiert dass
jedes Dockerfile ARG+ENV hat und jeder compose-build den Arg durchreicht.
Faellt bei jedem PR/Push gegen master, der einen neuen Service oder
Dockerfile ohne BUILD_SHA einfuehrt.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Pin-free pymdownx gets latest version which fixes NoneType error
on bare code fences in Python 3.11.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Older pymdown-extensions (10.12) crashes on bare code fences.
Upgraded to 10.14.3 + mkdocs-material 9.6.14.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
pymdownx.highlight requires language specification on code fences.
Bare ``` causes NoneType error during MkDocs build.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
There is only one remote (origin). Removed all occurrences of:
- git push gitea / git push origin main && git push gitea main
- "Pushing to gitea (external)" in deploy.sh
- # gitea: git@gitea.meghsakha.com:... remote comment in docs-src/index.md
- "Push auf gitea triggert" → "Push auf origin triggert" in docs
- Clone URL updated to ssh://git@coolify.meghsakha.com:22222/... in
README.md and CONTRIBUTING.md
Web UI URLs (gitea.meghsakha.com/...) are unchanged — those are still valid.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Squash of branch refactor/phase0-guardrails-and-models-split — 4 commits,
81 files, 173/173 pytest green, OpenAPI contract preserved (360 paths /
484 operations).
## Phase 0 — Architecture guardrails
Three defense-in-depth layers to keep the architecture rules enforced
regardless of who opens Claude Code in this repo:
1. .claude/settings.json PreToolUse hook on Write/Edit blocks any file
that would exceed the 500-line hard cap. Auto-loads in every Claude
session in this repo.
2. scripts/githooks/pre-commit (install via scripts/install-hooks.sh)
enforces the LOC cap locally, freezes migrations/ without
[migration-approved], and protects guardrail files without
[guardrail-change].
3. .gitea/workflows/ci.yaml gains loc-budget + guardrail-integrity +
sbom-scan (syft+grype) jobs, adds mypy --strict for the new Python
packages (compliance/{services,repositories,domain,schemas}), and
tsc --noEmit for admin-compliance + developer-portal.
Per-language conventions documented in AGENTS.python.md, AGENTS.go.md,
AGENTS.typescript.md at the repo root — layering, tooling, and explicit
"what you may NOT do" lists. Root CLAUDE.md is prepended with the six
non-negotiable rules. Each of the 10 services gets a README.md.
scripts/check-loc.sh enforces soft 300 / hard 500 and surfaces the
current baseline of 205 hard + 161 soft violations so Phases 1-4 can
drain it incrementally. CI gates only CHANGED files in PRs so the
legacy baseline does not block unrelated work.
## Deprecation sweep
47 files. Pydantic V1 regex= -> pattern= (2 sites), class Config ->
ConfigDict in source_policy_router.py (schemas.py intentionally skipped;
it is the Phase 1 Step 3 split target). datetime.utcnow() ->
datetime.now(timezone.utc) everywhere including SQLAlchemy default=
callables. All DB columns already declare timezone=True, so this is a
latent-bug fix at the Python side, not a schema change.
DeprecationWarning count dropped from 158 to 35.
## Phase 1 Step 1 — Contract test harness
tests/contracts/test_openapi_baseline.py diffs the live FastAPI /openapi.json
against tests/contracts/openapi.baseline.json on every test run. Fails on
removed paths, removed status codes, or new required request body fields.
Regenerate only via tests/contracts/regenerate_baseline.py after a
consumer-updated contract change. This is the safety harness for all
subsequent refactor commits.
## Phase 1 Step 2 — models.py split (1466 -> 85 LOC shim)
compliance/db/models.py is decomposed into seven sibling aggregate modules
following the existing repo pattern (dsr_models.py, vvt_models.py, ...):
regulation_models.py (134) — Regulation, Requirement
control_models.py (279) — Control, Mapping, Evidence, Risk
ai_system_models.py (141) — AISystem, AuditExport
service_module_models.py (176) — ServiceModule, ModuleRegulation, ModuleRisk
audit_session_models.py (177) — AuditSession, AuditSignOff
isms_governance_models.py (323) — ISMSScope, Context, Policy, Objective, SoA
isms_audit_models.py (468) — Finding, CAPA, MgmtReview, InternalAudit,
AuditTrail, Readiness
models.py becomes an 85-line re-export shim in dependency order so
existing imports continue to work unchanged. Schema is byte-identical:
__tablename__, column definitions, relationship strings, back_populates,
cascade directives all preserved.
All new sibling files are under the 500-line hard cap; largest is
isms_audit_models.py at 468. No file in compliance/db/ now exceeds
the hard cap.
## Phase 1 Step 3 — infrastructure only
backend-compliance/compliance/{schemas,domain,repositories}/ packages
are created as landing zones with docstrings. compliance/domain/
exports DomainError / NotFoundError / ConflictError / ValidationError /
PermissionError — the base classes services will use to raise
domain-level errors instead of HTTPException.
PHASE1_RUNBOOK.md at backend-compliance/PHASE1_RUNBOOK.md documents
the nine-step execution plan for Phase 1: snapshot baseline,
characterization tests, split models.py (this commit), split schemas.py
(next), extract services, extract repositories, mypy --strict, coverage.
## Verification
backend-compliance/.venv-phase1: uv python install 3.12 + pip -r requirements.txt
PYTHONPATH=. pytest compliance/tests/ tests/contracts/
-> 173 passed, 0 failed, 35 warnings, OpenAPI 360/484 unchanged
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Neue Endpunkte POST /obligations/dedup und GET /obligations/dedup-stats.
Pro candidate_id wird der aelteste Eintrag behalten, alle weiteren erhalten
release_state='duplicate' mit merged_into_id + quality_flags fuer Traceability.
Detail-View filtert Duplikate aus. MKDocs aktualisiert.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
DB-Constraint erlaubt nur must/should/may. 'can' gibt es nicht.
Alle Referenzen auf 'can' durch 'may' ersetzt.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Benjamin Admin
Benjamin_Boenisch