llm-tool-maker 0.2.0

An intelligent tool-making package that leverages LLMs to analyze projects and generate tools

LLM Tool Maker

An intelligent tool-making system that uses local LLMs (via Ollama) to analyze projects, generate tools, and execute them in a sandboxed environment — all through a modern web dashboard or CLI.

Features

• Project Analysis — AST-based scanning of project structure, imports, and entry points
• LLM Tool Generation — generates Python tools from natural language descriptions using local LLMs
• Sandboxed Execution — subprocess runner with module whitelist, no network, and configurable timeout
• Autonomous Pipeline — 6-stage DB-backed pipeline: Analyse → Plan → Validate → Implement → Test → Review, with auto-fix loop
• Web Dashboard — 6-page glass-morphism UI (Dashboard, Pipeline, Execute, Analyze, Provider, Config, Docs)
• Dependency Management — auto-detects imports, maps to pip packages, auto-installs via uv
• PostgreSQL Persistence — saves tools, plans, executions, and reviews; migration system included

Installation

# Install globally
uv tool install llm-tool-maker

# Or add to a project
uv add llm-tool-maker

From source

git clone https://github.com/codewithwest/project_tool-maker.git
cd project_tool-maker
uv sync

Quick Start

1. Start Ollama

ollama pull llama3.2  # or any model you prefer

2. Launch the dashboard

tool-maker ui

Opens at http://localhost:5000.

3. Or use the CLI

# Analyze a project
tool-maker analyze /path/to/project

# Run the autonomous pipeline
tool-maker pipeline "Create a function to parse CSV files"

# Execute a saved tool
tool-maker run my-tool

Database Setup

By default, tools work in-memory. For persistence and the full pipeline, set up PostgreSQL:

# Install PostgreSQL (Ubuntu/Debian)
sudo apt install postgresql
sudo systemctl start postgresql

# Create a database
sudo -u postgres createdb tool_maker

# Set the connection string (or export it)
export DATABASE_URL="postgresql://postgres:postgres@localhost:5432/tool_maker"

Then start the dashboard — migrations run automatically on first connection.

Environment Variables

Variable	Default	Description
DATABASE_URL	postgresql://postgres:postgres@localhost:5432/tool_maker	PostgreSQL connection string
OLLAMA_BASE_URL	http://localhost:11434	Ollama server URL
OLLAMA_MODEL	llama3.2	Default LLM model
SANDBOX_TIMEOUT	30	Tool execution timeout in seconds
MAX_FIX_ATTEMPTS	3	Auto-fix loop retry limit
TOOL_MAKER_CONFIG	~/.config/tool-maker/config.toml	Config file path

Web Dashboard

Pages

• Dashboard — overview, live terminal, quick stats
• Pipeline — run the 6-stage autonomous pipeline with progress tracking
• Execute — edit and run tools in an IDE-style editor; browse saved tools in the sidebar
• Analyze — scan a project and inspect its structure
• Provider — configure Ollama model and test prompts
• Config — manage database, migrations, sandbox whitelist, dependency approvals
• Docs — view release notes and this README, rendered as formatted markdown

API Endpoints

20+ REST endpoints powering the dashboard:

Method	Endpoint	Description
GET	/api/pipeline	Run the autonomous pipeline
GET	/api/analyze	Analyze a project path
POST	/api/execute	Execute a tool
POST	/api/fix	Fix a broken tool via LLM
POST	/api/refine	Refine a tool with instructions
POST	/api/explain	Get an LLM explanation of execution
GET	/api/deps/check	Check tool dependencies
POST	/api/deps/approve	Approve a package for auto-install
GET	/api/db/migrations	List migration status
POST	/api/db/migrate	Run pending migrations

Python API

from tool_maker import ToolMaker

tm = ToolMaker(
    llm_provider="ollama",
    base_url="http://localhost:11434",
    model="llama3.2"
)

# Analyze a project
info = tm.analyze_project("/path/to/project")

# Generate and execute a tool
result = tm.create_and_execute_tool("Parse CSV files and return row count")
print(result)

# Or run the full pipeline
result = tm.run_pipeline("Build a CLI tool that counts lines of code")
print(result)

Project Structure

src/tool_maker/
├── __init__.py          # ToolMaker orchestrator
├── analyzer/             # AST-based project scanner
├── cli/                  # CLI argument parsing and handlers
├── config.py             # Configuration loading
├── db/                   # PostgreSQL models, connection, migrations
├── llm/                  # Ollama provider (HTTPX-based)
├── planner/              # Planner, validator, executor, reviewer
├── tool/                 # Generator, executor, sandbox, fixer, deps
└── ui/                   # Flask web app (routes, templates, static)

Development

uv sync
uv run ruff check .
uv run pytest

License

MIT