AI Optimizer MCP 🧠🔧 - Multi-Task MCP Server

Développer par Barack Ndenga ♥️

Détails

Multi-tâche MCP Server pour VSCode/Cursor, CLI, agents autonomes. Optimisation code IA + tests + extensible.

Transports: Stdio (VSCode), subprocess, HTTP (futur)
Use Cases: VSCode chat, agent loops, CI/CD, remote servers
Sécurité: Env vars, sandbox exec

Manifeste (Multi-Task Capabilities)

🛠️ 3+ Tools: Code test/optimize/objective (+extensibles)
🔌 VSCode/Cursor: mcp.json natif
🖥️ CLI Standalone: ai-optimizer-mcp run
🤖 Agents: examples/agent.py loop
⚙️ Multi-Env: Local/dev/prod via .env
📊 Memory/History: JSON persistent
🔄 Boucles Itératives: Auto-improve

Configuration Multi-Plateforme

1. VSCode/Cursor (Recommandé)

Fichier .vscode/mcp.json (multi-servers):

{
  "servers": {
    "ai-optimizer": {
      "command": "python",
      "args": ["-m", "ai_optimizer_mcp.server"]
    },
    "ai-optimizer-dev": {
      "command": "python",
      "args": ["-m", "ai_optimizer_mcp.cli", "run", "--dev"]
    }
  }
}

Multi-task: Switch servers en chat!

2. CLI / Scripts / Agents

ai-optimizer-mcp run  # Stdio server (pipes)
ai-optimizer-mcp run --dev  # Debug
ai-optimizer-mcp --install-mcp  # Print mcp.json

3. Agents Autonomes / Subprocess

# examples/agent.py
import asyncio
from mcp.client.stdio import stdio_client

async def agent_loop():
    async with stdio_client(command=["python", "-m", "ai_optimizer_mcp.server"]) as client:
        # Multi-task calls
        score = await client.call_tool("run_tests", {"code_snippet": code})
        improved = await client.call_tool("generate_improvement", {"code": code, "test_result": score})

Prérequis (.env)

cp .env.example .env
# OPENAI_API_KEY=sk-...
# OBJECTIVE="Your custom goal"

Usage Multi-Tâche

VSCode Chat: use_mcp_tool("ai-optimizer", "run_tests", ...)
CLI Pipe: echo code | ai-optimizer-mcp run
Agent Loop: python examples/agent.py
CI/CD: Subprocess dans GitHub Actions/Jenkins

Exemple Réponse Tool:

run_tests → "Tests passed: score=4/4 (f(2)=4)"
generate_improvement → "def f(x): return 2 * x"

Troubleshooting Multi-Env

VSCode: Reload window après mcp.json
No API Key: ValueError → Check .env
Timeout: TEST_TIMEOUT=10 in .env
Memory: rm memory.json
Logs: --dev ou LOG_LEVEL=DEBUG

Développement

pip install -e .[dev]
pre-commit install
pytest

Tools MCP (Extensibles)

| Tool | Args | Use Case | |--------------------------|---------------------------|---------------------------| | run_tests | code_snippet: str | VSCode/CLI test code | | generate_improvement | code, test_result | Auto-optimize | | get_objective | - | Read goal any context |

Apache 2.0 - Multi-task ready! VSCode, CLI, Agents, CI. Contribute!

CHANGELOG

MCP Servers