Updated today
by timwuhaotian
killi
Open-source AI coding agent for your terminal, written in Rust. One 7 MB native binary for Apple Silicon Macs, ~18 MB RAM, 70–90% prompt-cache hits. Works with OpenAI, Anthropic and any compatible LLM; MCP, subagents, plan mode, skills, hooks, AGENTS.md. A lightweight Claude Code / Codex CLI alternative.
52 GitHub stars (52)
+52 in 21 days
| Period | New stars |
|---|---|
| Nov 11 | +3 |
- 2 skills (2 skills found in its SKILL.md files)
- 10 forks (10 forks)
- Today last update (Last updated today)
Install
Skills CLI
npx skills add timwuhaotian/killi
Asks which skills and agents to install. Just one: npx skills add timwuhaotian/killi --skill refresh-metrics — add -g for every project.
Claude Code
git clone https://github.com/timwuhaotian/killicp -r "killi/.agents/skills/refresh-metrics" ~/.claude/skills/
Copies the refresh-metrics skill into ~/.claude/skills/ for every project (use .claude/skills/ for one project). Repeat for the other skills listed below.
Plugin
/plugin marketplace add timwuhaotian/killi
Run inside Claude Code if the repository ships a plugin marketplace, then pick plugins from /plugin.
Skills can run code on your machine — read the SKILL.md first.
What it does
In the author's own words, from each skill's SKILL.md.
-
refresh-metrics
Re-measure killi's installed size, startup, memory and first-request tokens against Kimi Code and Claude Code, regenerate the docs/bench-*.svg, docs/eval-*.svg and combined docs/benchmarks-*.svg charts, and update the numbers README.md quotes. Use after a release, when Kimi Code or Claude Code updates, or when asked to refresh the benchmark charts or the size, memory and token claims.
.agents/skills/ on GitHub (opens in a new tab)refresh-metrics/ SKILL.md -
self-eval-improve
Evaluate and improve killi from its own session logs in ~/.killi/sessions. Two modes: review (read-only — measure recent real use, read the log behind each lead, confirm provider behavior with a probe, report ranked findings) and improve (the full loop — attribute each loss, land one narrow fix, verify it live both ways, record the decision, ship a patch release when asked). Use when asked how ki…
.agents/skills/ on GitHub (opens in a new tab)self-eval-improve/ SKILL.md
No skill matches that filter.
Review before you install
A skill is a set of instructions and scripts your agent will follow, and it can run code on your machine. Read the SKILL.md and any scripts first — this listing is not a security review.