userid prioritizedly obtained from the request header of the mcp client

[Hchunjun]

Refactor user ID extraction to use AsyncLocalStorage for better context management.

📋 Description

🔄 Type of Change

• 🐛 Bug fix (non-breaking change which fixes an issue)
• ✨ New feature (non-breaking change which adds functionality)
• 💥 Breaking change (fix or feature that would cause existing functionality to not work as expected)
• 📚 Documentation update
• 🎨 Style/UI changes
• ♻️ Code refactoring
• ⚡ Performance improvements
• 🧪 Test updates
• 🔧 Build/CI changes

🧪 Testing

• I have tested this change locally
• I have added tests that prove my fix is effective or that my feature works
• New and existing unit tests pass locally with my changes

📱 Screenshots (if applicable)

🔍 Code Review Checklist

• Code follows the project's coding standards
• Self-review of the code has been performed
• Code is properly commented, particularly in hard-to-understand areas
• Changes generate no new warnings
• Any dependent changes have been merged and published

📚 Related Issues

🚀 Deployment Notes

📋 Additional Context

[nullure]

Please test it locally, if it passes please let me know.

[Hchunjun]

Passed testing: Locally running the backend service, the Claude code is configured as follows for mcp; it can achieve priority by using the mcp client to specify the request header x-user-id to obtain the uid, avoiding the uncertainty issue of passing the uid when using mcp with AI

{
"headers": {
"Accept": "application/json, text/event-stream",
"Content-Type": "application/json",
"x-api-key": "oooo",
"x-user-id": "xxxx"
},
"type": "http",
"url": "http://localhost:8080/mcp"
}

[nullure]

MCP supports many ide's, from the looks of it. Your code might only support claude code, please make it flexible

[ajitam]

so to confirm right now I'm defining MCP (in VS Code) like this

 "mem": {
      "url": "http://dockint.lan:8080/mcp",
      "type": "http",
      "headers": {
        "x-api-key": "xyz...="
      }
    }

and then I would just change it to

 "mem": {
      "url": "http://dockint.lan:8080/mcp",
      "type": "http",
      "headers": {
        "x-api-key": "xyz...="
        "x-user-id": "my-user-id"
      }
    }

and it would overwrite any user_id which would be set my LLM in tool calling?

That would be great, I just setup a "shared" server and I was asking my self how can this be done. When LLM posted memory with user_id=null it just decided and used something for id - how do I define this otherwise?

@nullure isn't this JSON standard? As far as I understand all major clients will accept this (together with additional header) - but then again - I could be wrong.