ИИ-агенты
agent-qa
A self-improving QA agent that writes tests in plain English.
★ 902 звездыNOASSERTION
Смотреть руководство →prime-agent — проект с открытым исходным кодом из категории «ИИ-агенты» от PrimeIntellect-ai (PrimeIntellect-ai/prime-agent на GitHub). У него 21 619 звёзд на GitHub и 2 385 форков, он распространяется по лицензии NOASSERTION и последний раз обновлялся на GitHub 8 октября 2026.
A self-improving agent for coding workflows and long-running autonomous tasks
Pick a coding or research task that benefits from an agent that runs and improves over hours.
Install it with the official install script for your platform.
Launch the agent in a project and let the daemon keep sessions alive.
Give it a task, detach, and reattach later to review progress and results.
A persistent Python REPL is the built-in model tool - files, shell, tools and context are all managed through code.
Agents can spawn real child agents for parallel or background work and consume their results programmatically.
A refine flow reviews the trajectory and applies small, evidence-backed updates to harness state, with snapshots for rollback.
Daemon-backed agents keep running after the terminal disconnects and can be reattached, resumed and inspected.
Skills are importable Python packages, and a built-in creator can turn recurring workflows into new skills.
Running agents exchange messages and orchestrate one another directly.
Get the source code from GitHub and enter the project directory.
git clone https://github.com/PrimeIntellect-ai/prime-agent.git
cd prime-agentUse the CLI to browse, attach to and inspect running agents and the background service.
prime-agent agents
prime-agent status
prime-agent attach <agent>On Windows, the README provides a PowerShell installer instead of the shell script.
# Run the PowerShell install command from the README's Install sectionIt is maintained by Prime Intellect (PrimeIntellect-ai) on GitHub.
It is open source - see the LICENSE file in the repository for the exact terms (GitHub metadata lists the license as NOASSERTION).
The agent drives an LLM, so you need model access - a provider account or API key you configure - and usage costs depend on that provider.
Compared with single-session coding agents, Prime Agent is built around durability: background daemon sessions, persistent state and self-refinement for tasks that run for hours, not minutes.
Install with the script in the README's Install section, open a project, start the agent, and use prime-agent agents and attach to follow it.
A self-improving QA agent that writes tests in plain English.
Recursive self-improving physical agent OS — verified, evidence-based execution for robots and simulators
Agent harness for multi-agent swarms: deploy, coordinate and federate autonomous workflows
Yohei Nakajima's autonomous task agent - now a self-building agent framework in Python
PraisonAI 🦞 — Hire a 24/7 AI Workforce. Stop writing boilerplate and start shipping autonomous self-improving agents that research, plan, code, and execute t…
Self-managing, unbounded context and long-term memory for coding agents: one session, for life.
Другие проекты из категории «ИИ-агенты» с наибольшим числом звёзд в каталоге LibreAI — реалистичный список альтернатив, если prime-agent вам не подходит.
A personal AI that really does things — on any OS, with your data staying yours.
An agentic skills framework & software development methodology that works.
Agent harness optimization system with skills, memory and security for Claude Code, Codex and more
The agent that grows with you — a Python framework for adaptive AI agents.
DeepSeek's plugin-based agent harness where everything is a plugin.
Соседние проекты из категории «ИИ-агенты» в каталоге, ближайшие по числу звёзд к prime-agent — продолжайте изучать полку по обе стороны от него.
A visual all-in-one platform for building, debugging, and deploying AI agents.
The veteran open-source framework for conversational AI that you host.
Google's code-first toolkit for building real AI agents.
Give it a GitHub issue and it tries to fix the code for you - the original SWE-bench agent
Swarm — OpenAI's experimental framework for lightweight multi-agent orchestration with agents and handoffs, MIT
Kortix - the open-source AI operating system: agents, skills and memory as files in a repo
Бесплатно и открыто — забирайте исходный код, ставьте звезду репозиторию и начинайте создавать.