首要代理：一個自我改進的RLM框架

摘要

語言模型是序列處理器，但長時程代理需要模型權重與活動上下文之外的外部資訊與計算。Prime Agent 是一個開源框架，用於長時程評估與編碼代理工作流程。一個常駐的 IPython REPL 遵循遞迴語言模型抽象，以進行程式化上下文處理與測試時計算，而 Continual Harness 跨軌跡保留歷史記錄、記憶、技能、提示詞與子代理規格。遞迴子代理透過代理對代理的直接通訊來協調，而代理視圖（Agents View）讓人類可以檢視與管理由 daemon 支援的會話。Prime Agent 標準化了執行、復原、驗證與資源核算，同時將策略構建留給模型。這種低摩擦、高表達力的介面可防止框架失敗演變成模型失敗，並將測量推向模型真實的最大潛在能力。Prime Agent 將 ARC-AGI-3 RHAE Best@1 從 30% 提升至 95.5%，並在長上下文編程、GPU 核心生成、模擬器建構與自主 nanoGPT 速通方面，匹配或超越原生及主流框架。在 Factorio 中，我們發現精煉能實現持續的科技進展，而專用子代理能實現並行化工作。程式碼位於 https://github.com/PrimeIntellect-ai/prime-agent。

English

Language models are sequential processors, but long-horizon agency requires external information and computation beyond model weights and active context. Prime Agent is an open-source harness for long-horizon evaluation and coding-agent workflows. A persistent IPython REPL follows the Recursive Language Model abstraction for programmatic context processing and test-time compute, while Continual Harness preserves histories, memories, skills, prompts, and subagent specifications across trajectories. Recursive subagents coordinate through direct agent-to-agent communication, and the Agents View lets humans inspect and manage daemon-backed sessions. Prime Agent standardizes execution, recovery, verification, and resource accounting while leaving strategy construction to the model. This low-friction, expressive membrane prevents harness failures from becoming model failures and pushes measurement toward the model's true maximal underlying capability. Prime Agent raises ARC-AGI-3 RHAE Best@1 from 30% to 95.5% and matches or exceeds native and popular harnesses across long-context coding, GPU-kernel generation, emulator construction, and autonomous nanoGPT speedruns. On Factorio, we find refinement allows for continuous technology progression and dedicated subagents enable parallelized work. Code is available at https://github.com/PrimeIntellect-ai/prime-agent.