MCP server exposing the Lextiva compliance reference dataset — 29 privacy + AI regulations (incl. LGPD, Quebec Law 25, Swiss FADP, 18 US state laws), jurisdictions, document types, data-subject rights, and the matrix between them. Lets Claude / Cursor / any MCP-aware AI agent answer privacy and compliance questions with verifiable, primary-sourced data.
Strong domain-specific tool set with clear naming (verb_noun pattern: list_*, get_*, documents_for_*, etc.) and comprehensive descriptions (avg 180 chars, well above 10-char minimum). All 12 tools have input schemas with type definitions. However, parameter descriptions are minimal or absent in several tools (e.g., list_regulations, list_jurisdictions accept empty input but lack explanation of what they return). Output schemas are documented in code but not formally declared in tool registration. Error handling is present and actionable (e.g., 'Unknown regulation id... Use list_regulations to see all valid ids'). No tool annotations (readOnlyHint, destructiveHint) despite all tools being read-only. Composition is excellent, tools chain naturally (list → get → cross-product queries). No security concerns (read-only, no secrets, no state mutation).
Given a jurisdiction id, returns every regulation that applies there, plus the required and optional documents for each. Use this when a user asks 'I serve customers in California — what privacy laws apply?'.
Returns the specific required-clauses checklist for a (regulation, document_type) pair, plus primary-source URLs and contextual notes. Most useful tool for 'what do I need in my Privacy Policy under GDPR?' style queries.
Given a regulation id, returns every document type that regulation requires (or optionally permits), with the required clauses for each. Use this when a user asks 'what documents do I need for X?'.
Returns the full record for a single data subject right. Pass ids like 'erasure', 'portability', 'opt_out_sale_or_sharing'.
Returns the full record for a single document type — typical sections, audience, and which regulations mandate it. Pass ids like 'privacy_policy', 'dpa', 'ai_use_policy'.
List tools (list_regulations, list_jurisdictions, list_document_types, list_data_subject_rights) have empty input schemas but lack descriptions explaining what they return or when to call them.
No tool annotations (readOnlyHint, destructiveHint, idempotentHint) despite all 12 tools being read-only and idempotent. Adding readOnlyHint=true to all tools would signal to agents that these are safe to call without side effects.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | C | 64 | 2026-07-28+ | v2 |
Returns the full record for a single jurisdiction including the list of regulations that apply there and population context. Pass ids like 'eu', 'uk', 'us', 'us_ca', 'us_va', 'ca'.
Returns the full record for a single regulation including key_articles, scope, applicable_to, data_subject_rights granted, supervisory authority, and primary-source URL. Pass the short id like 'gdpr', 'ccpa_cpra', 'eu_ai_act', 'vcdpa'.
Returns the list of the 8 canonical data-subject rights and the regulations that grant each: access, rectification, erasure, portability, restriction, objection, opt_out_sale_or_sharing, automated_decision_making.
Returns the list of the 7 compliance documents tracked: privacy_policy, terms_of_service, cookie_policy, gdpr_notice, ccpa_notice, ai_use_policy, dpa. Each carries audience, mandatory_under (which regulations require it), and typical_sections.
Returns the list of all jurisdictions tracked: countries, supranational unions, and US states. Currently 21 jurisdictions. Each carries ISO code, type (country/us_state/supranational), and the regulation IDs that apply there.
Returns the list of all privacy and AI regulations covered, with short_name, full name, jurisdiction, effective_date, and official_url for each. Currently 29 regulations across EU, UK, US (federal + 10 states), and Canada. Use this first when the user asks about coverage or when you need to look up a regulation ID for follow-up calls.
Given a document type id, returns every regulation that mandates or permits that document, with required clauses for each. Use this when a user asks 'which laws require a Privacy Policy / DPA / Cookie Policy?'.
Output schemas are documented in code comments and return statements but not formally declared in tool registration. LLMs cannot see the structure of returned fields (e.g., regulation.key_articles, regulation.scope) without explicit schema documentation in the tool definition.
No pagination support on list_* tools. If regulations, jurisdictions, or document types grow beyond ~50 items, returning all at once risks context window exhaustion. Add optional limit and offset/cursor parameters.
Error messages reference tool names (e.g., 'Use list_regulations to see all valid ids') but do not provide structured error codes or categorization (retryable vs user-fixable). Agents cannot programmatically distinguish error types.