LongerAgentExploring agent autonomy.

A TUI demo that gives the coding agent tools to manage its own context and workflow.

Agent-Driven Context

The agent can inspect its own context distribution and proactively compress selected blocks to keep the session productive over long tasks.

Spawn explorer and executor agents that run concurrently within the same session, each with their own context window, and report back when done.

Type messages at any time — even mid-task. Messages are queued and delivered at the next activation boundary. No need to wait, no need to restart.

Anthropic, OpenAI, Kimi, MiniMax, GLM, Ollama, oMLX, LM Studio, and OpenRouter. Cloud or local, your choice.

Install, manage, and create reusable skill packages. The built-in skill-manager teaches the agent to find and install new skills autonomously.

AGENTS.md files and the Important Log survive across sessions and context resets. The agent remembers what matters.