Skip to content

goSprinto/compliance-skills / gdpr-compliance-checker

gdpr-compliance-checker

Verified

Autonomous GDPR compliance auditor that scans a codebase to identify PII collection, storage, and sharing, then produces an article-by-article gap analysis, a pre-filled Data Processing Agreement (DPA), and a ROPA (Record of Processing Activities) starter kit — all exported as downloadable files (.docx recommended). Use this skill whenever the user mentions GDPR, data privacy audit, DPA, ROPA, EU compliance, PII scan, data protection, ICO, CNIL, or asks whether their codebase or product is compliant with data protection law. Trigger even for casual phrasings like "are we GDPR compliant?", "I need to check our data privacy", or "help me get audit-ready." The skill runs autonomously — it reads the codebase, researches third-party processors, and produces final deliverables without the user needing to provide structured input.

references/sector-overlays.md

View on GitHub
# Sector-Specific Overlays

GDPR intersects with sector regulations that impose stricter or additional requirements. Identify the company's sector from the codebase and load the relevant sections.

**How to identify sector from codebase**:
- Health fields, MDR references, prescription data → Healthcare
- Payment processing, KYC fields, AML references, PSD2 → Financial services
- Age gates, parental consent, child-friendly UI → Children's services
- Employee records, payroll, time tracking → HR tech
- Location data at scale, smart devices, sensor data → IoT / smart infrastructure
- AI model training on personal data, automated decisions → AI / ML

---

## Table of Contents
1. [Healthcare and health tech](#1-healthcare-and-health-tech)
2. [Financial services and fintech](#2-financial-services-and-fintech)
3. [Children's services](#3-childrens-services)
4. [HR tech and employment](#4-hr-tech-and-employment)
5. [Advertising technology](#5-advertising-technology)
6. [AI and machine learning](#6-ai-and-machine-learning)
7. [Telecommunications](#7-telecommunications)
8. [Sector gap analysis additions](#8-sector-gap-analysis-additions)

---

## 1. Healthcare and Health Tech

### Relevant regulations (alongside GDPR)
- **EU Medical Device Regulation (MDR 2017/745)** — software as a medical device (SaMD)
- **EU In Vitro Diagnostic Regulation (IVDR 2017/746)**
- **Clinical Trials Regulation (EU No 536/2014)**
- **France**: Hébergeur de Données de Santé (HDS) certification required for hosting health data
- **Germany**: Digital Health Applications (DiGA) regulation for prescription apps
- **UK**: Data Security and Protection Toolkit (DSPT) for NHS-related processing

### Key additional requirements

**Data classification**: Health data is special category data (Art. 9). Processing requires:
- Explicit consent (Art. 9(2)(a)), OR
- Healthcare provision (Art. 9(2)(h)) + professional secrecy obligation, OR
- Public health (Art. 9(2)(i)), OR
- Research (Art. 9(2)(j)) + appropriate safeguards

**DPIA mandatory**: Large-scale health data processing always requires a DPIA (Art. 35(3)(b)).

**Security standards**: Health data demands higher security standards. Check for:
- Encryption at rest (AES-256 minimum)
- Role-based access (clinicians vs admin vs billing)
- Audit trails on all access to patient records
- Pseudonymisation of research data

**France (HDS)**: Any entity hosting health data (including cloud infrastructure for a health app) must be HDS-certified or use an HDS-certified host. Major certified hosts: AWS (with French healthcare option), Microsoft Azure France, OVHcloud. Non-certified hosting = serious violation.

**Retention**: Medical records have statutory retention periods that override standard GDPR minimisation — e.g. UK NHS records: 10 years post last treatment; varies significantly by country and record type.

**Gap signals**:
- Health fields detected in schema (diagnosis, medication, health_status, prescription)
- No DPIA found for health data processing
- No mention of HDS certification (France)
- No mention of professional secrecy obligations
- Health data shared with third-party analytics

---

## 2. Financial Services and Fintech

### Relevant regulations
- **PSD2 (Payment Services Directive 2)** — open banking, strong customer authentication
- **MiFID II** — financial instrument data, record keeping
- **AML Directive (AMLD)** — anti-money laundering, KYC, transaction monitoring
- **DORA (Digital Operational Resilience Act)** — ICT risk management from Jan 2025
- **UK FCA rules** — Consumer Duty, financial promotions, data handling

### Key additional requirements

**KYC / AML data retention**: AML regulations require retention of customer due diligence (KYC) data for **5 years** after the end of the business relationship. This overrides GDPR minimisation — retention is a legal obligation.

**Transaction records (MiFID II)**: Must retain transaction records for **5 years** (certain records 7 years). These are legal obligations.

**PSD2 consent**: Open banking access to payment accounts requires explicit, specific consent. Third-party providers (TPPs) must have clear authorisation.

**Strong Customer Authentication (SCA)**: PSD2 requires SCA for electronic payments — biometric or two-factor. Gap signal: payment flow without SCA.

**DORA (from Jan 2025)**:
- ICT risk management framework required
- Major ICT incident reporting (within 4 hours initial notification, 72 hours intermediate, 1 month final)
- ICT third-party risk management (assess all critical ICT providers)
- Digital operational resilience testing

**Credit data**: Strict rules on credit scoring and automated credit decisions — must inform applicant of automated decision, provide human review on request (overlaps Art. 22 GDPR).

**BNPL (Buy Now Pay Later)**: Increasingly regulated — consumer credit rules may apply; data about creditworthiness is sensitive.

**Gap signals**:
- payment, card_number, iban, sort_code fields without encryption at rest
- No 5-year retention for KYC/AML data
- Automated credit/risk decisions without Art. 22 compliance
- PSD2 open banking access without specific consent
- Post-DORA: no ICT incident response plan

---

## 3. Children's Services

### Relevant regulations
- **GDPR Art. 8** — consent for children
- **EU Better Internet for Kids / EDPB guidance on children**
- **UK Age Appropriate Design Code (Children's Code)** — applies to any service "likely to be accessed" by under-18s
- **US COPPA** — if any US children's data processed (note: outside GDPR scope but relevant for global products)

### Key requirements

**UK Children's Code (15 standards)**:
1. Best interests of the child — default to high privacy settings
2. Data protection impact assessments — DPIA required
3. Age appropriate application — tailor privacy information to child's age
4. Transparency — clear, child-appropriate privacy information
5. Detrimental use of data — no use of children's data in ways detrimental to their wellbeing
6. Policies and community standards — uphold stated policies
7. Default settings — privacy settings default to high
8. Data minimisation — collect minimum necessary
9. Data sharing — no data sharing beyond what is necessary
10. Geolocation — off by default
11. Parental controls — provide tools for parental oversight
12. Profiling — off by default; only on with compelling reason
13. Nudge techniques — no nudges that encourage children to share more data
14. Connected toys and devices — meet the Code's standards
15. Online tools — provide obvious, accessible privacy controls

**Gap signals**:
- No age verification or age estimation
- Geolocation on by default
- Profiling features without separate consent
- Nudge UX patterns (e.g. "Your friends have shared their location!")
- No child-appropriate privacy notice
- Parental consent mechanism absent

---

## 4. HR Tech and Employment

### Relevant regulations
- **GDPR Art. 88** — member states may provide more specific rules for employment context
- **National employment law** — varies significantly; see member-state-supplements.md
- **Germany**: §26 BDSG, works council requirements
- **France**: Labour Code employee monitoring rules
- **Spain**: Digital rights in employment (LOPDGDD Arts. 87–91)

### Key requirements

**Legal basis for employee data**:
- Performance management, payroll, benefits: Contract (Art. 6(1)(b))
- Legal obligations (tax, social security): Legal obligation (Art. 6(1)(c))
- Occupational health: Separate legal basis under Art. 9(2)(b) + national law
- Monitoring: Legitimate interests — but requires strict necessity test; many national laws add co-determination requirements

**Consent in employment**:
- Consent is generally **not valid** as a legal basis for employee data processing because of the power imbalance — employees cannot freely refuse
- Exception: genuinely optional benefits (e.g. optional health screening programme)

**Employee monitoring**:
- Email/internet monitoring: Must be proportionate, transparent, and notified in advance; works council/union consultation required in many states
- Location tracking: Strict necessity test; not permitted for home workers in most states without specific justification
- Keystroke/screen monitoring: Highly invasive; difficult to justify
- BYOD: Personal device policies create complexity — work data vs personal data

**HR data retention**:
- Recruitment data (unsuccessful candidates): typically 6 months–1 year max
- Employee records (current): duration of employment + statutory minimum
- Payroll records: typically 7 years (tax obligations)
- Disciplinary records: typically 12 months after completion of proceedings
- Time and attendance: typically 3–5 years

**Gap signals**:
- Employee monitoring features without transparency notice or works council reference
- Location tracking of remote employees
- Candidate data retained beyond 6 months without explicit consent
- No separate policy for employment data processing
- Consent used as legal basis for employment processing

---

## 5. Advertising Technology

### Relevant regulations
- **ePrivacy Directive** — consent for tracking (see eprivacy-checklist.md)
- **IAB TCF** — challenged in Belgium (GBA decision 2022); under scrutiny across EU
- **Digital Services Act (DSA)** — large platforms; targeting restrictions

### Key requirements

**Real-time bidding (RTB)**:
- Sharing user IDs and profile data in RTB auctions requires GDPR-compliant consent — high risk
- Belgian DPA found IAB TCF's consent mechanism systematically non-compliant (2022) — significant if TCF is used

**Profiling for advertising**:
- Profiling using special category data (health, political views, etc.) prohibited without explicit consent
- Profiling of children for advertising: prohibited

**Digital Services Act (DSA) — very large platforms (>45M EU users)**:
- Must provide option to use service without targeted advertising
- Cannot target based on sensitive attributes (health, religion, sexuality, political views)
- Must provide transparent information about recommendation systems

**Gap signals**:
- ad_profile, interest_segments, lookalike_audience fields
- User profiling for ad targeting without Art. 9 check on inferred sensitive categories
- IAB TCF integration without awareness of Belgian ruling
- No opt-out from profiling for advertising

---

## 6. AI and Machine Learning

### Relevant regulations
- **EU AI Act** — fully applicable from 2026 (tiered by risk)
- **GDPR Art. 22** — automated decision-making
- **EDPB Guidelines on automated decision-making**

### Key requirements

**EU AI Act risk tiers**:
- **Unacceptable risk** (prohibited): Real-time biometric surveillance in public spaces, social scoring by governments, subliminal manipulation
- **High risk**: Biometric categorisation, critical infrastructure management, education/employment decisions, essential services access, law enforcement, migration management, administration of justice
- **Limited risk**: Chatbots, deepfakes — transparency obligations only
- **Minimal risk**: Spam filters, AI in games — no obligations

**GDPR Art. 22 — automated decision-making**:
- Automated decisions with legal or significant effects require: human oversight option, ability to contest, explanation of the decision
- Profiling that infers sensitive attributes (even without explicit data) = Art. 9 risk

**Training data**:
- Training ML models on personal data = processing personal data; requires legal basis
- Legitimate interests commonly used — but requires LIA
- Data minimisation: consider pseudonymisation or synthetic data alternatives

**Gap signals**:
- ML model training on user data without documented legal basis
- Automated recommendations or scoring with significant effects on users
- No human review option for automated decisions
- Biometric data used in ML (face recognition, voice ID)
- Model infers sensitive attributes (political views, health, sexuality)

---

## 7. Telecommunications

### Relevant regulations
- **ePrivacy Directive** — most directly applicable to telecoms
- **Specific national telecoms data retention laws** — many EU states require mandatory retention of traffic data for law enforcement

### Key requirements

**Traffic and location data**:
- Call records, SMS metadata, IP connection data: may only be processed for billing, quality of service, security — not for profiling without consent
- Location data requires consent or strict necessity
- Law enforcement data retention: national laws vary post-CJEU rulings limiting blanket retention

---

## 8. Sector Gap Analysis Additions

After the standard gap analysis, add a **Sector-Specific Findings** section:

```
SECTOR IDENTIFIED: [healthcare / fintech / children's / HR / adtech / AI / other]
ADDITIONAL REGULATIONS TRIGGERED:
  - [Regulation]: [specific requirement] → [Severity] → [Finding] → [Recommendation]

SECTOR RISK SUMMARY:
  Critical sector gaps: N
  High sector gaps: N
  Sector compliance posture: [Not Started / Partial / Mostly Compliant / Audit-Ready]
```
rea-analysis-optimizationmorluto/reaOptimize REA providers and shared analysis workflows for agent context, latency, CPU, and memory while preserving correctness and Evidence. Use for analysis bottlenecks and cache or query design; CI/test-suite optimization belongs to rea-ci-tests-audit.Stars: 472kUpdated Claude CodeCodexVerifiedInstall
rea-ci-tests-auditmorluto/reaAudit REA CI and test suites for correctness, fixture drift, false passes, and measured performance; use for suite or workflow investigations and efficiency changes, rather than ordinary feature testing or a single test failure.Stars: 472kUpdated Claude CodeCodexVerifiedInstall
rea-tool-designmorluto/reaAudit, investigate, design, or change REA investigation tools, CLI/MCP contracts, provider capabilities, and Evidence semantics.Stars: 472kUpdated Claude CodeCodexVerifiedInstall
ponytail-auditDietrichGebert/ponytailQuality audit of the whole repo: bugs, security, real load, missing tests, speed, and what to delete. Most important first.Stars: 196kUpdated Claude CodeCodexVerifiedInstall