TEMP MCP v1.1.0

Released: 2025-12-17
Location: /opt/mcp-servers/temp/mcp_temp_server.py

Overview

Version 1.1.0 updates the TEMP MCP server for multi-tenant operation. All tools now require a user parameter for proper isolation and delegate AI compatibility.

Changes from v1.0.0

User Parameter Required

All tools now require explicit user parameter:

stage - Stage content for processing

{
  "user": "chris",
  "content": "# Research Notes\n...",
  "source_type": "research"
}

get - Retrieve staged content

{
  "user": "chris",
  "temp_id": "tmp_abc123"
}

list - List staged items

{
  "user": "chris"
}

status - Check TEMP environment status

{
  "user": "chris"
}

cleanup - Clean up specific staged item

{
  "user": "chris",
  "temp_id": "tmp_abc123"
}

cleanup_expired - Clean up all expired items

{
  "user": "chris"
}

Implementation Details

async def temp_stage(self, args: Dict[str, Any]) -> Dict[str, Any]:
    user = args.get("user") or self.default_user

    if not user:
        return {"content": [{"type": "text", 
                "text": "❌ 'user' required for multi-tenant operation"}]}

Workflow Integration

TEMP staging is designed for the research → PDF workflow:

1. Stage - Store research/content with 24h TTL
2. Process - Use temp_id with docs.create_pdf
3. Cleanup - Automatic via auto_cleanup or manual

temp.stage(user="chris", content="...") → tmp_xyz
docs.create_pdf(user="chris", temp_id="tmp_xyz", auto_cleanup=true)
→ PDF created, temp content removed

Storage Location

Staged content stored in user-specific directories:

/data/temp/{user}/tmp_{short_id}/
  ├── content.json
  ├── metadata.json
  └── [optional files]