Add MCP (Model Context Protocol) support to prompt registry microservice

[dittops]

Description

This PR adds Model Context Protocol (MCP) support to the prompt registry microservice, enabling AI agents to discover and use prompt management capabilities through the OPEA MCP framework.

Changes Made

1. MCP Integration
   - Added MCP support to the prompt registry service with three exposed tools:
   • create_prompt: Store prompts with user association
   • get_prompt: Retrieve prompts by user, ID, or keyword search
   • delete_prompt: Delete prompts by ID
     - Enabled via ENABLE_MCP environment variable
2. Service Architecture Updates
   - Fixed service registration pattern to use unified service name opea_service@prompt_registry for all endpoints
   - Added ServiceType.PROMPT_REGISTRY to the constants enum
   - Updated service startup logic in the main block
3. Documentation
   - Updated README files to explain MCP functionality and usage
   - Documented important limitation: When MCP is enabled, regular HTTP endpoints are not available (MCP-only mode)
   - Added examples of MCP client usage
4. Testing
   - Added comprehensive test script (test_prompt_registry_mcp.sh) that tests both MCP-enabled and disabled scenarios
   - Created unit tests for MCP functionality (test_prompt_registry_mcp_unit.py)
   - Tests verify backward compatibility and proper MCP tool registration
5. Dependencies
   - Added mcp library to requirements

Technical Details

• When ENABLE_MCP=true, the service operates in MCP-only mode, exposing only the SSE endpoint (/sse) for MCP protocol communication
• Regular HTTP endpoints (/v1/prompt/create, /v1/prompt/get, /v1/prompt/delete) remain available when MCP is disabled
• MCP tools are named after the function names, following the framework's convention

Testing

The PR includes comprehensive tests that verify:

• Backward compatibility with MCP disabled
• MCP tool discovery and initialization
• SSE endpoint availability
• Proper service registration with unified naming

[chensuyue]

Wait for CI machine back for test.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

Files with missing lines	Coverage Δ
comps/cores/mega/constants.py	97.36% <100.00%> (+0.03%)	⬆️

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[joshuayao]

Hi @dittops could you please help check the CI failures? Thanks.

[letonghan]

lgtm