Prime Agent

Prime Agent is a local AI coding and research agent for developers and researchers whose tasks need to continue across terminal sessions. It runs commands and edits files in your project directory, while background sessions keep working after you disconnect. The code is open source under the MIT License. Model access uses a subscription or an API-key provider.

Its main distinction is persistent working state. A Python control environment retains context alongside memories, reusable skills, and instructions for subagents. Its Recursive Language Model approach lets the agent manage context and call tools through code, which suits extended coding tasks and research evaluations.

Prime Agent can review a session and save focused lessons as supplemental instructions or memories. These refinements stay local to the session by default, leave the base system prompt intact, and have recorded snapshots for rollback. A built-in skill creator can turn recurring workflows into executable Python skills for a project or personal use.

Subagents can work in parallel or in the background, exchange messages directly, and remain available across turns. Persistent goals track an objective's progress; schedules and periodic wakeups let a session return to work later. Autonomous runs have turn, token, and time budgets and can check user-defined quality gates. The agent executes generated code with your user permissions; its worker processes aren't a security sandbox.