Repolode
Предложить
PrimeIntellect-ai

PrimeIntellect-ai/prime-agent

An open-source coding and research agent built on a persistent Python REPL, with subagents, a self-refining harness, and support for long-running tasks.

Open source альтернатива:Claude CodeDevin

Открыть на GitHub
Звёзды
22,161
Форки
2,451
Открытые issues
146
Последний коммит
3 часа назад

Что это?

Prime Agent is an open-source coding and research agent for general and long-running work. Its built-in model tool is a persistent Python REPL, so file operations, shell commands, tool calls, subagents, and context management all run as code. Child agents can be spawned with rlm.spawn(...) for parallel or background work and return their results programmatically.

The project is built around two ideas: a Recursive Language Model approach that treats context as variables and tools as function calls, and a Continual Harness that stores supplemental prompts, memories, skill descriptions, and reusable subagent specifications as durable state. The /refine command reviews the current trajectory and applies small, evidence-backed updates to that state, with snapshots that support rollback. The immutable base system prompt is never rewritten.

Sessions can run as daemon-backed background agents that survive terminal disconnects and can be reattached later. Features such as persistent goals, heartbeats, schedules, and bounded autonomous mode support long tasks, and an optional Slack connector is available through MCP.

Ключевые возможности

  • 01Persistent Python REPL as the main tool for files, shell, and subagents
  • 02Built-in subagents via rlm.spawn for parallel or background work
  • 03/refine updates supplemental harness state with snapshots for rollback
  • 04Daemon-backed sessions that keep running after the terminal disconnects
  • 05Persistent goals, heartbeats, schedules, and bounded autonomous mode
  • 06Executable skills packaged as importable Python modules

Сценарии использования

  • Running long coding tasks that continue across terminal sessions
  • Parallelizing research or evaluation work across child agents
  • Turning recurring workflows into reusable project or personal skills
  • Running scheduled or periodic checks on an ongoing project

Темы

#ai-agent#coding-agent#llm#python-repl#subagents#rust#autonomous-agent#developer-tools

Частые вопросы

Что такое prime-agent?
An open-source coding and research agent built on a persistent Python REPL, with subagents, a self-refining harness, and support for long-running tasks. Prime Agent is an open-source coding and research agent for general and long-running work. Its built-in model tool is a persistent Python REPL, so file operations, shell commands, tool calls, subagents, and context management all run as code. Child agents can be spawned with rlm.spawn(...) for parallel or background work and return their results programmatically.
prime-agent можно использовать бесплатно? Какая у него лицензия?
prime-agent распространяется с открытым исходным кодом под лицензией Other. Перед использованием в коммерческих продуктах проверьте условия лицензии.
prime-agent всё ещё поддерживается?
Активно поддерживается. Последний коммит в prime-agent — 19 окт. 2026 г..
Для чего можно использовать prime-agent?
Running long coding tasks that continue across terminal sessions Parallelizing research or evaluation work across child agents Turning recurring workflows into reusable project or personal skills Running scheduled or periodic checks on an ongoing project
Где скачать prime-agent?
Исходный код и инструкции по установке — на GitHub: https://github.com/PrimeIntellect-ai/prime-agent

Описание создано ИИ на основе README проекта. Статистика GitHub обновляется ежедневно.

Похожие проекты