bub
Bub is a deliberately small Python agent runtime built on agents.md and the Agent Skills format. Its defining idea is that every stage of a turn, from resolve_session and load_state through build_prompt and run_model to render_outbound and dispatch_outbound, is a pluggy hook. Builtins are registered first and external plugins load after them, so you can override one stage or replace the whole flow without forking the runtime.
It started in group chats, where several humans and agents share one conversation, so it avoids hidden session state: context is rebuilt from append-only tape records that you can inspect and replay. The same inbound pipeline runs across the CLI, Telegram and any channel you add, with adapters changing only the surface.
Install is a one-line script (or uv sync from source), and the CLI covers interactive chat, one-shot runs and a gateway listener mode. Configuration is environment variables (model, API key, step and token limits, spill threshold). It ships CLI, Telegram, tools, skills and model execution in the core, all replaceable. License is Apache-2.0.
Features
Why choose it
Trade-offs
Maintenance check
- Last commit
- 2026-10-19
- Commits, last 90 days
- 90
- Releases, last 90 days
- 9
- Latest release
- 0.6.0a3 (2026-10-18)
- Open issues
- 2
- Recent issues with at least one comment
- 2/4
Oldest week on the left; the latest four weeks are orange. Source: GitHub commit activity.
Status rule: archived on GitHub = Archived; last commit within 90 days = Active; 91 to 365 days = Slowing; older = Stale. All figures come from the GitHub API on the date shown.
Most discussed open issues
- #355 拆出 core 使 bub 更加 meta0 comments · 2026-10-12
Issue titles are quoted from GitHub. Open the link to check whether a problem still applies to the version you use.
Compatibility
Quick start
Use cases
Alternatives
Comments
No comments yet. Be the first!