Loop Engineering Python Packages

ouroboros-ai

Agent OS: the agent gets smarter on its own. We just hold the line: Interview-gated, staged evaluation, budgeted evolution loop. MCP server, 14 runtimes: Claude Code, Codex CLI, Gemini CLI, OpenCode, Copilot, Kiro and more.

agent-toolkit-cli

🛠️ Composable AI agent toolkit — skills, agents, loops, and MCP templates for Claude Code, Cursor, OpenCode, Copilot, Windsurf, and Pi

qwenpaw

Your Personal AI Assistant; easy to install, deploy on your own machine or on the cloud; supports multiple chat apps with easily extensible capabilities.

entroly-core

Cut AI context cost without trusting the compressor. Every reduction is reversible, byte-exact recoverable, and carries an auditable receipt. Local-first, works through proxy, MCP, SDK, or agent wrapper.

loopx

A control plane with a durable state kernel for long-horizon agents and teams. Keep work moving and improving across sessions, with less human attention.

agentbundle

The complete AI operating model for software teams — from first idea to production. Three peer-supervised loops (discovery → build → release) over a catalogue of curated packs: skills, subagents, and hooks, each installed in one line. It's npm for your coding agent. Any agent, any stack — Claude Code, Codex, Cursor, Copilot, Gemini, Kiro.

entroly

Cut AI context cost without trusting the compressor. Every reduction is reversible, byte-exact recoverable, and carries an auditable receipt. Local-first, works through proxy, MCP, SDK, or agent wrapper.

dotbrain

Keeps your project's context private and your code repo clean

memu-py

Personal memory across agents

dyro

DyroEngineeringFlow 是面向多仓团队的本地优先工程自动化与交付控制平台。它将开发线、Git worktree、任务编排、 Agent 启动、质量门禁、独立复核与合并审计统一到可版本化配置中，帮助团队从任务创建到交付合并形成清晰、安全、 可追溯的自动化流程。日常通过简洁命令 dyro 即可完成初始化、环境检查、任务执行与协作交付。

millrace-ai

Turns repeatable, domain-agnostic workflows into graph-driven loops.

bounded-loops

Bounded Loops & Graphs: 69 agent loops (65 keyless), 7 reference graphs, independent gates, declared bounds, verifiable receipts. Paper: arXiv:2609.27871

lh-harness

The long-horizon computer-use harness. Run AI agents across desktop apps and the CLI for extended periods while preserving task state and making reliable progress on complex workflows. Features fresh-context execution, durable verified state, independent auditing, recoverable progress, and native Claude Code / Codex / OpenClaw integration.

lhht

The long-horizon computer-use harness. Run AI agents across desktop apps and the CLI for extended periods while preserving task state and making reliable progress on complex workflows. Features fresh-context execution, durable verified state, independent auditing, recoverable progress, and native Claude Code / Codex / OpenClaw integration.

memu-cli

Personal memory across agents

bumblehive

An agent runtime designed for tool-driven loop engineering.

tvastar

A durable Python harness for agents that act on real systems.

credbroker

The complete AI operating model for software teams — from first idea to production. Three peer-supervised loops (discovery → build → release) over a catalogue of curated packs: skills, subagents, and hooks, each installed in one line. It's npm for your coding agent. Any agent, any stack — Claude Code, Codex, Cursor, Copilot, Gemini, Kiro.

coremind-ai

配置驱动、可恢复的 Agent Runtime：CLI/TUI、TypeScript/Python SDK、Protocol v2、宿主验收、有界 Loop、上下文压缩、Replay、Child Run 与安全执行环境

5dive

Run a company of AI agents on a server you own. Spin up named agents (claude, codex, pi…), put them on an org chart with a shared backlog, let them hand off work and ping your phone only when a human must decide. MIT.

qzagent

Your Personal AI Assistant; easy to install, deploy on your own machine or on the cloud; supports multiple chat apps with easily extensible capabilities.

le-loopctl

Unified CLI for Loop Engineering — validate, score, trace, pipeline, agents, bench

loopgym

OpenAI Gym for LSS-defined agent loops

harborrag

Enterprise knowledge infrastructure for RAG and agents. Every answer is traceable to immutable evidence and checked against the live version before it is ever returned.