claude-skill-data-governance
novapulseaibe/claude-skill-data-governance
Claude Code skill for data governance — NIS2, SOC2, ISO 27001 compliance automation
Instalação
npx convez-ai novapulseaibe/claude-skill-data-governance1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.
Métricas
README
A tradução para o português deste README ainda está na fila.
/data-governance — Claude Code Skill
A Claude Code skill that enforces data governance best practices on any AI-generated project. Run it before you write a single line of code to establish governance foundations, or on an existing project to score compliance and surface gaps.
Compliance frameworks enforced automatically: NIS2 (Directive 2022/2555), SOC2 (AICPA Trust Service Criteria), ISO/IEC 27001:2022.
Quick Start
Open Claude Code in any project directory and type:
/data-governance
That's it. The skill detects whether you're starting fresh or auditing an existing project, and takes it from there.
Two Modes
Pre-audit mode — start here, before coding
Triggers automatically when the directory is empty or contains no source files.
The skill asks you 16 short questions (one at a time) covering ownership, data classification, authentication, compliance scope, and team contacts. Based on your answers, it generates 6 governance documents and commits them to git.
/data-governance # auto-detects empty directory → pre-audit
/data-governance --pre-audit # force pre-audit on an existing directory
What you answer:
| # | Question | Example answer |
|---|---|---|
| 1 | Project name | billing-api |
| 2 | One-sentence description | REST API that handles invoice generation and payment processing |
| 3 | Owner name & email | Jane Doe / jane@example.com |
| 4 | Team | Platform |
| 5 | Change approver | Tech Lead |
| 6 | Data classification | confidential |
| 7 | Processes personal data (GDPR)? | yes |
| 8 | What personal data? | customer names, email addresses, payment metadata |
| 9 | Retention period | 7 years (legal requirement) |
| 10 | Authentication mechanism | JWT |
| 11 | Externally facing? | yes |
| 12 | Environments | dev, staging, prod |
| 13 | Primary language | EN |
| 14 | Coding standards link | (skip) |
| 15 | Additional regulations | PCI-DSS |
What gets generated:
| File | Contents |
|---|---|
GOVERNANCE.md | Full governance charter — ownership table, escalation path, data classification, compliance scope, risk register, decision log |
SECURITY.md | Security policy — authentication, access control, secrets management, encryption, incident response procedure, NIS2/SOC2/ISO 27001 checklists |
DATA-MANAGEMENT.md | Data inventory, GDPR data subject rights, retention & disposal table, encryption policy, backup policy |
CONTRIBUTING.md | Branch strategy, development workflow (TDD), commit convention, PR checklist, security requirements |
ONBOARDING.md | Day-one guide — access setup, local setup, key documents table, compliance responsibilities |
README.md | Project README (only if none exists) — description, setup, environments table, links to governance docs |
.env.example | Secrets template with placeholder keys (created if absent) |
.gitignore | .env added if not already excluded |
Everything is committed in a single chore: add governance artifacts (NIS2/SOC2/ISO27001) commit.
Audit mode — scan an existing project
Triggers automatically when the directory contains source files.
The skill scans your project, scores each of the 5 governance pillars against NIS2, SOC2, and ISO 27001 controls, and produces a compliance report.
/data-governance # auto-detects source files → audit
/data-governance --audit # force audit mode
What gets scanned:
- Presence and content of
GOVERNANCE.md,README.md,SECURITY.md,DATA-MANAGEMENT.md,CONTRIBUTING.md,ONBOARDING.md,CHANGELOG.md,.env.example,.gitignore - Source files for hardcoded secrets (
password =,api_key =,token =, etc.) .github/workflows/for CI/CD presence and security scanning steps (audit,snyk,trivy,semgrep)- Project-type-specific checks (see Supported Project Types)
What gets generated:
GOVERNANCE-REPORT.md — a full compliance report containing:
# Governance Audit Report — billing-api
Global compliance score: 62/100
## Pillar Scores
| Pillar | Weight | Score | Status |
|-------------------------------|--------|--------|------------|
| Gouvernance organisationnelle | 30% | 45/100 | ❌ non-compliant |
| Structuration documentaire | 20% | 80/100 | ✅ compliant |
| Gestion des données | 20% | 60/100 | ⚠️ partial |
| Sécurité et accès | 20% | 55/100 | ⚠️ partial |
| Culture et formation | 10% | 90/100 | ✅ compliant |
## Gaps by Pillar
### Gouvernance organisationnelle — 45/100 ❌
| Severity | Control ID | Control | Framework | Fix |
|----------|--------------|--------------------------------------|-----------|--------------------------------------------------|
| critical | NIS2-ART21-1 | Risk management policy documented | NIS2 | Add a Risk Management section to GOVERNANCE.md |
| critical | A.5.1 | Information security policies defined| ISO 27001 | Create GOVERNANCE.md and SECURITY.md |
| major | CC1.2 | Management oversight defined | SOC2 | Add change approval process to GOVERNANCE.md |
## Critical Gaps Summary
- [NIS2-ART21-1] Risk management policy documented — Fix: Add Risk Management section to GOVERNANCE.md
- [A.5.1] Information security policies defined — Fix: Create GOVERNANCE.md and SECURITY.md
...
The report is committed as docs: add governance audit report.
Console summary:
📊 Audit complete — global score: 62/100
Critical gaps: 4 | Major: 7 | Minor: 2
Full report: GOVERNANCE-REPORT.md
Pillar Scoring
The global score is a weighted average of 5 pillars:
| Pillar | Weight | What it checks |
|---|---|---|
| Gouvernance organisationnelle | 30% | Owner defined, escalation path, risk management, data classification, incident reporting |
| Structuration documentaire | 20% | README complete, CHANGELOG present, operating procedures documented |
| Gestion des données | 20% | Data inventory, GDPR compliance, retention policy, encryption documented |
| Sécurité et accès | 20% | No hardcoded secrets, auth documented, access control, SAST/dependency scanning |
| Culture et formation | 10% | CONTRIBUTING.md, ONBOARDING.md, security training referenced |
Score thresholds:
- ✅ ≥ 80 — Compliant
- ⚠️ 50–79 — Partial — remediation required before production
- ❌ < 50 — Non-compliant — do not deploy
Supported Project Types
The skill auto-detects project type from manifest files and applies additional checks:
| Type | Detection signals | Extra checks |
|---|---|---|
web-app | React/Vue/Angular/Svelte/Next/Nuxt in package.json | CSP, XSS, HTTPS, CSRF |
api | Express/FastAPI/Gin/NestJS, routes/ or openapi.yaml | Auth, rate limiting, input validation, versioning |
service | Dockerfile, docker-compose.yml, k8s dir | Container hardening, secret injection, health check |
cli | bin field in package.json, Cobra/Click/Typer | Credential storage, shell injection |
library | main/exports in package.json, no server entry | API docs, CHANGELOG, semver policy |
data-pipeline | Airflow/Dagster/dbt/Prefect configs | Data lineage, PII in pipeline, data quality |
ml-model | *.ipynb, torch/sklearn/transformers in requirements | Training data provenance, PII removal, model card, artefact exclusion from git |
Compliance Coverage
Every audit maps findings to specific controls:
NIS2 (Directive 2022/2555, Art. 21 & 23)
- Risk management, incident handling, supply chain security, access control, encryption, MFA, vulnerability management, incident reporting SLA
SOC2 (AICPA 2017 Trust Service Criteria)
- CC1–CC8 (Security), P1–P7 (Privacy): integrity, oversight, logical access, vulnerability management, change management, data purpose limitation
ISO/IEC 27001:2022 (Annex A)
- A.5 Organisational controls, A.6 HR controls, A.8 Technological controls: classification, cryptography, access management, secure SDLC, security testing, configuration management
Workflow Integration
Recommended workflow for new AI-generated projects:
1. AI generates your project scaffold
2. /data-governance ← pre-audit: governance foundations
3. Start coding
4. /data-governance --audit ← before first production deploy
5. Fix critical gaps
6. /data-governance --audit ← verify remediation
Re-running audit: Safe to run multiple times. Each run produces a fresh GOVERNANCE-REPORT.md reflecting the current state.
Skill Location
The skill is installed globally and available in every project:
~/.claude/skills/data-governance/
SKILL.md ← agent instructions
templates/ ← Handlebars templates for generated files
checklists/ ← NIS2, SOC2, ISO 27001 control definitions
project-types/ ← Per-type extra checks
Files in This Repository
Skill - Data Governance/
README.md ← this file
docs/
superpowers/
specs/2026-06-04-data-governance-design.md ← design specification
plans/2026-06-04-data-governance-skill.md ← implementation plan