io.github.mdfifty50-boop/domain-expertise

domain-expertise-mcp

Inject structured professional domain expertise into AI agents. Accounting, Islamic finance, cybersecurity, real estate, HR, and 15+ more domains — all via MCP.

Why This Exists

AI agents are generalists. They know a little about everything but lack the structured reasoning patterns that professionals develop over decades. This MCP server bridges that gap — giving any agent access to decision frameworks, professional workflows, terminology, and regulatory context across 24 professional domains.

No other MCP server provides structured domain expertise injection.

What It Does

Core Tools

• get_domain_expertise — Get structured professional knowledge at three depth levels (overview, practitioner, expert). Includes decision frameworks, key principles, common pitfalls, and expert reasoning patterns.
• list_available_domains — See all 24 available professional domains with coverage details.
• get_decision_tree — Get step-by-step decision logic for specific professional scenarios (e.g., "Is this transaction Sharia-compliant?", "Should I register for VAT?").
• get_terminology — Professional terms with definitions in English and Arabic. Includes usage context and common misunderstandings.
• get_regulatory_context — Regulatory bodies, applicable laws, compliance requirements, and recent changes by jurisdiction.
• assess_complexity — Score a task's complexity and get recommendations for agent capability level needed.
• get_professional_workflow — Standard professional workflows with quality checkpoints and common shortcuts that cause problems.

Supported Domains (24)

Finance: Accounting, Tax, Audit, Islamic Finance, Conventional Banking, Investment, Insurance (Underwriting & Claims)

Legal: Corporate Law, Contract Law, Employment Law

Healthcare: Clinical, Administration, Pharmaceutical

Real Estate: Commercial, Residential

Technology: Cybersecurity, Data Privacy

Operations: Supply Chain, Logistics, Construction, Oil & Gas

People: Education, HR Management

Regional Coverage

Global, GCC, MENA, EU, US, UK, Asia — with GCC-specific knowledge for Saudi Arabia, UAE, Kuwait, Bahrain, Oman, and Qatar.

Installation

Claude Desktop / Claude Code

{
  "mcpServers": {
    "domain-expertise": {
      "command": "npx",
      "args": ["domain-expertise-mcp"]
    }
  }
}

Cursor

Add to .cursor/mcp.json:

{
  "mcpServers": {
    "domain-expertise": {
      "command": "npx",
      "args": ["domain-expertise-mcp"]
    }
  }
}

Use Cases

Building a financial agent? Inject accounting expertise so it handles revenue recognition, lease classification, and IFRS compliance like a CPA.

Islamic finance automation? Get Sharia compliance decision trees, Murabaha/Sukuk structures, and AAOIFI standards for your agent.

Cybersecurity tool? Add incident response workflows, NIST CSF alignment, and NCA/NESA compliance knowledge.

HR automation for GCC? Nitaqat/Saudization calculations, Emiratisation requirements, and end-of-service computations.

Example

Agent: "I need to determine if a client's Islamic finance product is Sharia-compliant"

→ get_decision_tree(domain: "islamic_finance", scenario: "sharia compliance")

Returns a structured decision tree:
1. Does the transaction involve riba (interest)? → If yes: NOT COMPLIANT
2. Is there excessive gharar (uncertainty)? → If yes: NOT COMPLIANT
3. Is the underlying activity halal? → If no: NOT COMPLIANT
4. Is risk genuinely shared? → If no: REVIEW NEEDED
5. If all pass → LIKELY COMPLIANT. Submit to Sharia Board.

Depth Levels

Pricing

Requirements

• Node.js 18+
• No API keys needed for free tier

License

MIT

Keywords

domain-expertise, mcp, mcp-server, professional-knowledge, ai-agent, accounting, islamic-finance, cybersecurity, legal, real-estate, healthcare, knowledge-injection, industry-knowledge, model-context-protocol