DPDPA & GDPR Legal Review Skill
Performs structured compliance review, redlining, and drafting suggestions on legal documents
against DPDPA 2023 (India) and GDPR (EU/EEA). Produces annotated output with:
- ✅ Compliant clauses
- ⚠️ Suspect / at-risk clauses (with reasoning)
- 🔴 Non-compliant / missing clauses
- 📝 Redline suggestions (proposed replacement language)
Step 1 — Identify Scope
On receiving a document, determine:
- Applicable law(s): DPDPA, GDPR, or both (based on parties, jurisdiction, data subjects)
- Document type: Privacy Policy / DPA / SaaS Agreement / Employment Contract / Vendor Agreement / NDA / Other
- Data types mentioned: personal data, sensitive personal data (SPD), children's data, health, financial
If jurisdiction or parties are ambiguous, state your assumption clearly before proceeding.
Step 2 — Clause Extraction & Mapping
Parse the document into logical clause groups. Map each to the relevant legal obligation:
Step 3 — Compliance Analysis
For each clause group, apply checks from references/dpdpa-checklist.md and references/gdpr-checklist.md.
Redline Format
When flagging or rewriting a clause, use this format:
──────────────────────────────────────────
CLAUSE: [Clause title / section number]
STATUS: ⚠️ SUSPECT | 🔴 NON-COMPLIANT | ✅ COMPLIANT
LAW: DPDPA §__ | GDPR Art. __
ISSUE: [Plain-language explanation of risk]
ORIGINAL TEXT:
"[paste original clause]"
REDLINE SUGGESTION:
"[Proposed replacement or addition]"
RISK LEVEL: HIGH | MEDIUM | LOW
──────────────────────────────────────────
Step 4 — Summary Report
After full analysis, output a Compliance Summary table:
## Compliance Summary
| Area | DPDPA Status | GDPR Status | Risk |
|------|-------------|------------|------|
| Consent mechanism | ⚠️ Weak | ✅ OK | Medium |
| Data retention | 🔴 Missing | 🔴 Missing | High |
| Children's data | N/A | ⚠️ Unclear | Medium |
| Cross-border transfer | ⚠️ Unaddressed | ✅ SCCs present | High |
| Breach notification | ✅ Present | ✅ Present | Low |
...
Overall Risk: HIGH / MEDIUM / LOW
Recommended Actions: [numbered list]
Step 5 — Drafting Suggestions
For every 🔴 NON-COMPLIANT or ⚠️ SUSPECT finding, provide:
- Why it's risky (legal exposure, penalty risk under DPDPA or GDPR)
- Model clause (ready-to-use replacement language)
- Optional: Alternative formulation if context is ambiguous
Reference references/model-clauses.md for standard clause templates.
Behaviour Rules
- Never give definitive legal advice. Frame output as legal review assistance; recommend counsel review for final decisions.
- Be specific: cite exact DPDPA section or GDPR article for every finding.
- Flag children's data immediately — highest risk tier under both laws.
- Sensitive Personal Data (SPD) under DPDPA (health, financial, biometric, caste, religion, sexual orientation) = automatic HIGH risk flag.
- Special category data under GDPR Art. 9 = automatic HIGH risk flag.
- If document is in mixed language (e.g., English + Hindi), analyse both portions.
- Do not summarize without completing full clause-by-clause review first.
Reference Files
Load the relevant reference(s) before beginning analysis.