compliance-os – skill Agent & Tự động hóa cho agent AI

Compliance OS — Meta-Orchestrator

Multi-framework compliance program orchestration. Four decisions, no per-framework deep-dive:

1. Which frameworks apply to this company? — framework_selector.py ranks the 12 supported frameworks against a company profile (industry, geography, AI use, medical, financial, headcount, customers, healthcare-PHI, NIS2 essential/important entity, US gov contractor) and returns applicable ones with dependency graph

2. How much do selected frameworks overlap? — cross_framework_mapper.py computes control-level overlap with confidence rating; outputs unified control matrix + evidence-reuse opportunities

3. What does a mock audit produce? — audit_simulator.py generates 8–15 finding scenarios with severity distribution matching IIA expectations + interview questions per control

4. What's the unified evidence checklist? — evidence_pool_generator.py consolidates evidence across enabled frameworks; outputs which artefact satisfies which controls across which frameworks

This skill is NOT a per-framework deep-dive. The per-framework skills (${CLAUDE_PLUGIN_ROOT}/vendor/ra-qm-team/iso42001-specialist/, compliance-team-eu-ai-act/, ${CLAUDE_PLUGIN_ROOT}/vendor/ra-qm-team/gdpr-dsgvo-expert/, etc.) do the operational work. Compliance OS orchestrates them.

This skill is NOT a substitute for binding legal advice. Cross-framework mappings reflect published guidance (ISO standards, regulations, EDPB/Commission guidance, IIA / AICPA professional standards). Novel cross-walks should be reviewed with counsel.

Keywords

compliance orchestration, multi-framework compliance, compliance OS, cross-framework mapping, control overlap, evidence pool, evidence reuse, audit simulation, mock audit, internal audit programme, GRC, governance risk compliance, framework selector, compliance program, integrated compliance, ISO 19011, IIA IPPF, AICPA AT-C, NIST CSF profile, multi-cert program, SOC 2 + ISO 27001, ISO 27001 + ISO 42001, ISO 13485 + MDR 745, AI Act + ISO 42001, GDPR + ISO 27001, compliance officer, compliance team workflow, certification readiness

Quick Start

# Decision A: Which frameworks apply for the company?
python scripts/framework_selector.py                          # embedded mid-stage AI SaaS sample
python scripts/framework_selector.py path/to/profile.json

# Decision B: Compute cross-framework overlap
python scripts/cross_framework_mapper.py                      # embedded ISO 27001 + SOC 2 sample
python scripts/cross_framework_mapper.py path/to/control_libs.json

# Decision C: Simulate an audit
python scripts/audit_simulator.py                             # embedded ISO 27001 sample
python scripts/audit_simulator.py path/to/audit_scope.json

# Decision D: Consolidate evidence checklist across frameworks
python scripts/evidence_pool_generator.py                     # embedded 3-framework sample
python scripts/evidence_pool_generator.py path/to/program.json

Key Questions (ask these first)

• Have you named every applicable framework? Forgetting one means rebuilding the audit program later. Run framework_selector.py with your profile.
• What's the most certificate / regulation your company already operates? That's your reuse anchor. Map every new framework against it.
• What's the audit calendar? A multi-framework program means surveillance audits stacked through the year — plan auditor independence + capacity.
• Where is evidence stored? Multi-framework programs collapse when evidence lives in one team's drive without an index. Run evidence_pool_generator.py to surface the reuse opportunities.
• What's the management-review cadence across frameworks? Each framework wants its own management review, but a single integrated review (per ISO Annex SL) typically satisfies all of them with one calendar slot.
• Who owns the meta-program? If no single accountable role, the program fragments.

Core Responsibilities

1. Framework Selection

The framework: company-profile JSON in → applicable-framework list out with dependency graph.

Deterministic logic:

• Medical device → ISO 13485 + ISO 14971 + (EU MDR 745 if EU market) + (FDA QSR if US market)
• Customer-facing AI → ISO 42001 + EU AI Act (if EU users) + GDPR (if personal data)
• B2B SaaS with enterprise customers → SOC 2 + ISO 27001 (often required for procurement)
• EU customers + personal data → GDPR mandatory
• Highly regulated industry (financial, health) → additional sectoral overlays

Run framework_selector.py to apply the decision rules.

2. Cross-Framework Control Mapping

The framework: for each selected framework, parse its control library; compute overlap with other selected frameworks.

Per merged-control output:

• Mapping confidence (HIGH / MEDIUM / LOW)
• Evidence-reuse opportunity (single artefact satisfies N controls)
• Per-framework citation
• Implementation guidance reusable across frameworks

Densest known overlap: ISO 27001 Annex A ↔ SOC 2 Trust Services Criteria — historically ~75% control coverage shared. Adding ISO 42001 brings AI-specific controls; adding GDPR brings privacy-specific.

Run cross_framework_mapper.py with framework control libraries.

3. Audit Simulation

The framework: generate a realistic mock internal audit per ISO 19011 + IIA IPPF standards.

Per audit output:

• 8–15 finding scenarios per ISO 19011 typical depth
• Severity distribution: ≥ 40% observations/OFI, ≤ 15% critical/major (IIA expectation for healthy programs)
• Interview questions per scoped control (3–5 questions per control)
• Document-review request list
• Walk-through requests where applicable

Run audit_simulator.py with framework + scope.

4. Evidence Pool

The framework: consolidate evidence requirements across enabled frameworks; identify reuse opportunities.

Output:

• Evidence artefact list (e.g., access-review log, supplier risk register, incident log)
• Per artefact: list of (framework, control) tuples it satisfies
• Reuse-leverage score (artefact A satisfies N controls across M frameworks)
• Acquisition cost estimate (effort to produce + maintain)

Run evidence_pool_generator.py with program config.

Workflows

Workflow 1: Program Bootstrap (multi-framework, 4–8 weeks)

Goal: stand up a compliance program covering 2–4 frameworks simultaneously.

# 1. Run framework selector with company profile
python scripts/framework_selector.py profile.json
# 2. For each applicable framework, identify the per-framework skill and run its gap analysis
# 3. Run cross-framework mapper to identify reuse opportunities
python scripts/cross_framework_mapper.py control_libs.json
# 4. Run evidence pool generator to consolidate
python scripts/evidence_pool_generator.py program.json
# 5. Cross-check with cs-compliance-officer agent
# 6. Output: prioritized program backlog with owners + dates

Workflow 2: Annual Audit Calendar (yearly)

Goal: plan internal audit cycles covering all applicable frameworks.

# 1. Refresh framework selector if profile changed
python scripts/framework_selector.py profile.json
# 2. For each framework, run its internal-audit-plan tool
#    (e.g., aims_audit_scheduler.py for ISO 42001; isms_audit_scheduler.py for ISO 27001)
# 3. Coordinate the audit calendar across frameworks (auditor independence + capacity)
# 4. Run audit simulator for each framework to prep auditors
python scripts/audit_simulator.py scope.json
# 5. Output: integrated audit calendar with owners + auditor assignments

Workflow 3: Pre-Certification Readiness (per new framework, 6–12 weeks)

Goal: prepare for an external certification audit.

# 1. Run gap analysis for the new framework
#    (ISO 42001: aims_gap_analyzer.py; ISO 27001: compliance_checker.py; SOC 2: gap_analyzer.py)
# 2. Run cross-framework mapper against already-certified frameworks
python scripts/cross_framework_mapper.py control_libs.json
# 3. Reuse evidence for HIGH-confidence mappings; build new for MEDIUM/LOW
# 4. Run audit simulator to dry-run the certification audit
python scripts/audit_simulator.py scope.json
# 5. Close remaining gaps before external auditor stage 1

Workflow 4: Evidence Pool Consolidation (quarterly)

Goal: keep the unified evidence pool fresh + reusable.

# 1. Refresh evidence pool generator
python scripts/evidence_pool_generator.py program.json
# 2. Identify HIGH-reuse-leverage artefacts (1 evidence -> 5+ controls)
# 3. Confirm evidence freshness (within retention requirement per framework)
# 4. Audit the evidence pool itself (no orphan controls, no stale evidence)

Output Standards

**Bottom Line:** [one sentence — what's the multi-framework picture + biggest reuse opportunity]
**The Decision:** [one of: framework-set | overlap-map | audit-plan | evidence-consolidation]
**The Evidence:** [framework names + control IDs from the tool, not adject

… (nội dung đầy đủ có trong gói tải về)