MCP Code Execution

Implement context-efficient MCP integrations using code execution patterns instead of direct tool calls.

Core Concept

Present MCP servers as code APIs on a filesystem. Load tool definitions on-demand, process data in execution environment, only return filtered results to context.

Quick Start

1. Generate Tool API from MCP Server

python scripts/mcp_generator.py --server-config servers.json --output ./mcp_tools

Creates:

mcp_tools/
├── google_drive/
│   ├── get_document.py
│   └── list_files.py
├── salesforce/
│   ├── update_record.py
│   └── query.py
└── client.py  # MCP client wrapper

2. Discover Tools Progressively

from scripts.tool_discovery import discover_tools, load_tool_definition

# List available servers
servers = discover_tools("./mcp_tools")
# ['google_drive', 'salesforce']

# Load only needed tool definitions
tool = load_tool_definition("./mcp_tools/google_drive/get_document.py")

3. Use Context-Efficient Patterns

import mcp_tools.google_drive as gdrive
import mcp_tools.salesforce as sf

# Filter data before returning to context
sheet = await gdrive.get_sheet("abc123")
pending = [r for r in sheet if r["Status"] == "pending"]
print(f"Found {len(pending)} pending orders")  # Only summary in context

# Chain operations without intermediate context pollution
doc = await gdrive.get_document("xyz789")
await sf.update_record("Lead", "00Q123", {"Notes": doc["content"]})
print("Document attached to lead")  # Only confirmation in context

Multi-Agent Workflow

For complex tasks, delegate to specialized sub-agents:

1. Discovery Agent: Explores available tools, returns relevant paths
2. Execution Agent: Writes and runs context-efficient code
3. Filtering Agent: Processes results, returns minimal context

See references/patterns.md for implementation details.

Tool Discovery Strategies

Filesystem Exploration

List ./mcp_tools/ directory, read specific tool files as needed.

Search-Based Discovery

from scripts.tool_discovery import search_tools

tools = search_tools("./mcp_tools", query="salesforce lead", detail="name_only")
# Returns: ['salesforce/query.py', 'salesforce/update_record.py']

Lazy Loading

Only read full tool definitions when about to use them.

Context Optimization

• Before: 150K tokens (all tool definitions + intermediate results)
• After: 2K tokens (only used tools + filtered results)
• Savings: 98.7%

Persisting Skills

Save working code as reusable functions:

# ./skills/extract_pending_orders.py
async def extract_pending_orders(sheet_id: str):
    sheet = await gdrive.get_sheet(sheet_id)
    return [r for r in sheet if r["Status"] == "pending"]

Privacy & Security

Data processed in execution environment stays there by default. Only explicitly logged/returned values enter context.

Advanced Patterns

See references/patterns.md for:

• Aggregation without context bloat
• Cross-source joins
• Polling loops
• Batch operations
• Error handling