Trammel

IronAdamant/Trammel

★ 1 stars Python Code/Dev Tools Updated 5d ago

Stdlib-only planning and execution harness for LLM-assisted coding: dependency-aware decomposition, beam strategy branching, incremental verification, failure constraints, and recipe memory. MCP server included.

View on GitHub → 🔍 Audit Wallet Slippage →

Quick Install

Copy the config for your editor. Some servers may need additional setup — check the README.

Add to claude_desktop_config.json:

{
  "mcpServers": {
    "trammel": {
      "command": "uvx",
      "args": [
        "trammel"
      ]
    }
  }
}

Or install with pip: pip install trammel

README Excerpt

**Trammel helps AI coding assistants plan, verify, and remember multi-step tasks.** Instead of letting your AI assistant wing it on complex changes, Trammel breaks goals into ordered steps, verifies each one works, learns from failures, and saves successful strategies for reuse. It works with **Claude Code**, **Cursor**, and any MCP-compatible editor — or standalone via Python and CLI.

Tools (12)

constraintscritical_path_lengthfailure_patternslayer_widthsmax_parallelismplansrecipe_filesrecipe_trigramsrecipesstepstrajectoriesusage_events

Trammel

Quick Install

README Excerpt

Tools (12)

Related Code/Dev Tools Servers