Code agents

Generating shell commands and code using natural language models (OpenAI ChatGPT).

CLI to manage the OpenAI to Codex Wrapper API server

Claude Code - An AI-powered code assistant with comprehensive tools

Multi-agent coordination daemon and tools for OpenCode

Open-source Agent SDK backed by OpenCode ACP (drop-in replacement for claude_agent_sdk)

One command launcher for running OpenCode with a local llama.cpp model.

Multi-tier local LLM stack: llama-swap + FastAPI auto-router + opencode wiring.

OpenFisca Rules as Code model for Paraguay.

OpenGPT 3.5/4 is a project aimed at providing practical and user-friendly APIs. The APIs allow for easy integration with various applications, making it simple for developers to incorporate the natural language processing capabilities of GPT into their projects.

Dynamic agent factory for OpenHands that generates micro-agents based on technology keywords

OpenThaiGPT focuses on developing a Thai Chatbot system to have capabilities equivalent to ChatGPT, as well as being able to connect to external systems and be able to retrieve data flexibly. Easily expandable and customizable and developed into Free open source software for everyone.

AI Agent system for OptiProfiler — advisor, debugger, and results interpreter. CLI: opagent (recommended) or optiprofiler-agent.

Agent Supervisor for AI coding agents — a state-aware cockpit for managing multiple Claude Code, Codex, and Shell sessions in parallel.

Multi-agent orchestrator skill for Claude Code

Minimal decoder-only LLM experiments trained from scratch with PyTorch.

OSCAR - GitHub-Specialized AI Coding Assistant (multi-provider: Gemini, OpenAI GPT-5.5/5.4, Claude 4.8/4.6, Groq, Ollama)

MCP server for communicating with other AI agents via file-based I/O

Erasure-coded storage backends with Merkle integrity verification for the ourochronos ecosystem

Session feedback loop for Claude Code -- capture, detect, report, enforce

Secure OVH API proxy + Claude Code skill for AI agents

A BDD-focused fork of opencode

Token-oriented object notation (TOON) for compact, reversible JSON encoding

Contract-first multi-agent software engineering. Contracts before code. Tests as law. Agents that can't cheat.

Framework for multiagent systems development in Python

Coding Assistance using ChatGPT from OpenAI

parag: interactive visualization of higher-order graphs in Python

Incorporate long strings painlessly, beautifully into Python code.

Parallel multi-agent pipeline for code analysis and transformation.

a lib to parse llm answer to code

PAT — an AI coding agent built from scratch.

An interactive agent for codebase modification using LLMs

Multi-provider lightweight local coding agent

A python debugger with OpenAI integrations

Compiles all readable files into a single PDF for easy sharing with LLMs.

Peek Agent Service - where Peek Apps can run their I/O code

RAG scaffolder + embedded library: 7 backbones incl. code-graph/R2R/Onyx, advanced retrieval (contextual, parent-doc, query-expansion+RRF, CRAG), scored advisor, SaaS Query API with auth + rate limit

Declarative, JSON-native AI agent framework. Define workflows as data, not code.

Autonomous research agent — persistent memory, experiment tracking, hypothesis-driven proposals, GPU run lifecycle

永續性知識管理系統：基於 Obsidian-NotebookLM-Claude Code 三層防幻覺 RAG 框架

Local Slack Web to Codex bridge for Bob.

ProgramEngineerGPT is an interactive command line tool that leverages the power of AI to assist developers with code comprehension, exploration, and generation. It serves as a virtual assistant that can analyze codebases, answer queries about code, and even help in setting up new coding projects.

RAG-enabled agent for code refactoring assistance

Python SDK facade for Pi Coding Agent sessions, services, resources, tools, and extensions

Sync OpenCode agents and skills from an Obsidian vault to workspace directories

Agent runtime primitives for pidevkit

Coding-agent implementation for pidevkit

Interactive coding agent CLI

Claude Code hook that auto-compiles PineScript files via TradingView API

A powerful RAG (Retrieval-Augmented Generation) system built with LangChain, designed as an MCP server for Cursor, VS Code, and other AI assistants

Trace capture for Claude Code sessions - sync to Pisama platform for analysis

Local-first AI project management agent — reads your repo, documents it, watches for changes, builds LLM prompts.

Python Open Coding Agent using DeepAgents and LangChain. Aims to be close to GH Copilot.

Claude Code CLI plugin for Pollinations.ai — generate images, text, audio, video. BYOP auth. 50+ models.

Multi-agent framework for code generation, API testing, and UI automation

Portainer container management — Docker environments, stacks, Kubernetes clusters, registries, users, and edge devices.

Give Claude Code enterprise-grade Power BI superpowers — community-driven, AI-powered BI automation

Powerline segments for Claude Code: subscription quota + tmux attention indicator.

PR-Agent aims to help efficiently review and handle pull requests, by providing AI feedback and suggestions.

QodoAI PR-Agent aims to help efficiently review and handle pull requests, by providing AI feedbacks and suggestions.

Comprehensive GitHub PR health scoring with AI-powered analysis

Python code transformers that mimic pragma compiler directives

A powerful toolkit for building, managing, and composing hierarchical prompts as code.

AI-powered DSA question solver with video explanations using Manim animations

A Firebase Firestore RAG system that bridges natural language prompts with structured database queries. It converts user input into intelligent retrieval operations, allowing developers to interact with Firestore using plain English instead of writing queries manually.

preLLM — One function for small LLM preprocessing before large LLM execution. Like litellm.completion() but with decomposition.

Prime is a powerful, modular, and highly extensible framework designed for developers building applications that require a seamless blend of performance, flexibility, and scalability. With Prime, you can harness the power of modern tools to craft cutting-edge solutions across industries.

ProcessFork integration for Claude Code — adds /snapshot, /fork, /merge slash commands.

ProcessFork memory adapter for CrewAI ≥0.130 — time-travel debugging across crew runs.

Product intelligence for JetBrains IDEs via ACP

This proj_flow extension aids with builds covered with github.com/mbits-libs/coveralls

Deterministic loop-driven Codex CLI orchestration for project delivery.

Shared memory foundation for AI development teams

Intelligent LLM Execution Layer for developer teams — anonymize code, route models, solve tickets.

Intelligent LLM Execution Layer for developer teams — anonymize code, route models, solve tickets.

pytest + observability infrastructure for AI agents. Evaluate the agents you build: multi-turn adversarial red teaming and artifact grading with jury based scoring. Observe the agents you use: intelligent risk screening and intent trajectories for coding agents (Claude Code, Cursor). LLM agnostic.

CLI coding assistant powered by your Ollama GPUs via PastaWater

Bridge OpenAI tools to Claude Code SDK

Py Code Agent - AI Coding Assistant with LiteLLM

A2A Gateway plugin for Py Code Agent - enables agent-to-agent communication via A2A v0.3.0 protocol

Git plugin for Py Code Agent - provides git status and repository information tools

Unified code search plugin for Py Code Agent — regex + AST structural search

MCP Gateway plugin for Py Code Agent - connect MCP servers and expose their tools

Oh My OpenCode plugin collection for Py Code Agent — multi-model orchestration, parallel sub-agents, intent gate, and discipline enforcement

Plan Orchestrator plugin for Py Code Agent - advanced task orchestration with dynamic resource allocation and priority scheduling

Search plugin for Py Code Agent - provides web search capabilities via DuckDuckGo

LLM7-powered generator that creates a minimal single-function Python package and runner from a natural-language spec.

Python Dev Kit Inside Code Interpreter for AI Agents - Jupyter-based code execution with multimodal support

MCP (Model Context Protocol) Server for py4agent - Provides code execution capabilities through Jupyter kernels

A lightweight agent harness built with Textual and a configurable multi-provider backend

Agent skill manager for the Agent Skills standard (agentskills.io)

Package defining the PyAgentSpec library for Agents and LLM fixed-flows abstractions.

PyAiAgent is a modern, fast (high-performance), async framework for building AI agents with pythonic code.

A programming framework for agentic AI. Proxy package for autogen-agentchat.

pyChatGPT GUI - is an open-source, low-code python GUI wrapper providing easy access and swift usage of Large Language Models (LLMs) such as ChatGPT, AutoGPT, LLaMa, GPT-J, and GPT4All with custom-data and pre-trained inferences.

Installable Claude Code workflow for Python projects

Use local Claude Code CLI as a Pydantic AI model provider with full support for structured responses, tools, and streaming

PydanticAI toolset for the Ejentum Reasoning Harness. EjentumToolset subclasses pydantic_ai.FunctionToolset and registers eight tools: four dynamic (reasoning, code, anti-deception, memory) plus four adaptive variants (adaptive-reasoning, adaptive-code, adaptive-anti-deception, adaptive-memory) that pre-fit the cognitive operation to the caller's task via an adapter LLM.

Boring, opinionated helpers for PydanticAI that are so simple you didn't want to even vibe code them. (Unofficial)

A code-first evaluation framework for testing LLM agents' ability to use MCP tools