菜单
Audit the fixed context overhead every session starts with — system prompt, plugins, skills, AGENTS.md — measured from the real wire transcript
默认使用会先检查来源的 Prompt；你也可以切换为直接命令，或下载本地副本。
决定是否安装前，请先阅读 SKILL.md，以及 SkillsMP 当前展示的配套文件。
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt；运行前请先检查来源。
npx skills add https://github.com/egorfedorov/kimi-context-optimizer --skill kco-overhead命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地？可下载 SkillsMP 当前能够提供的文件。
Analyze AGENTS.md files for token bloat and suggest optimizations
Generate a compact project map so Kimi understands the codebase without opening every file
Configure token budget limits, auto-compact settings, optional token pricing, and view current budget status (window-aware — reads the active model's context size from config.toml)
正在显示 SKILL.md
基于 SOC 职业分类
| name | kco-overhead |
| description | Audit the fixed context overhead every session starts with — system prompt, plugins, skills, AGENTS.md — measured from the real wire transcript |
| whenToUse | When the user runs /kco-overhead or asks what their baseline/session-startup context cost is |
Measure how many tokens every session of this project pays BEFORE any work happens — and where to cut.
Run:
node ${KIMI_SKILL_DIR}/../../src/overhead.js
The report shows:
/kco-agentsmd,
uninstalling unused plugins, pruning skill descriptions).Present the output to the user as-is (it is already formatted). If the report says no wire transcript was found, explain that the audit needs at least one completed exchange in a session for this project.
Key framing for the user: baseline overhead is paid in EVERY session, so a one-time trim repays itself continuously — it is usually the highest-leverage optimization available.