Have you ever actually seen the agent's context? What if you could fully control it?

A developer released kamchatka, a Rust-based terminal LLM agent that makes context, tools, permissions, and requests explicit, first-class state rather than hidden internals. The agent's context tab displays every item, its token cost, whether it will be included in the next request, and the reason if not, letting users move items between active, elided, and excluded states, pin content that compaction cannot touch, and inspect the exact request sent to the model. The project positions legibility over autonomy as its core design goal.

Most terminal LLM agents become black boxes when context fills up. You can't inspect what the model is carrying, remove specific items, or know why something was dropped. kamchatka https://github.com/ljedrz/nachalnik/tree/master/kamchatka is different: a Rust terminal agent where context, tools, permissions, and requests are explicit, first-class state. The context tab shows every item, its cost, whether it's in the next request, and why not. You can move items between active/elided/excluded, pin things compaction can't touch, and see the exact request. Most agents optimize for autonomy. kamchatka optimizes for legibility.