菜单
Show ken's measured impact as a compact scoreboard, from ken's own benchmark results in benchmarks/results/. One-shot display, not a persistent mode, and not a per-repo number. Trigger: /ken-gain, "ken gain", "what does ken save", "show ken impact", "ken scoreboard".
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/rajnandan1/ken --skill ken-gain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。
Thompson-mode discipline for any coding task: think first, build bottom-up, brute force until measured, rewrite over patch. Not for non-coding requests.
Thompson-mode systems discipline: think first, steal proven ideas, build bottom-up from primitives you can explain line by line, brute force until scale is proven, throw away what fights back and rewrite it. Supports intensity levels: lite, full (default), ultra. Use on ANY coding task: writing, adding, refactoring, fixing, reviewing, or designing code, and choosing libraries or dependencies. Also use whenever the user says "ken", "ken mode", "thompson mode", "brute force it", "rewrite it", or "throw it out and redo it". Do NOT use for non-coding requests (general knowledge, prose, translation, summaries, recipes).
Audit the whole repo for Thompson-mode violations. A ranked list of what to rewrite, delete, or take back into the trusted base.
正在显示 SKILL.md
| name | ken-gain |
| description | Show ken's measured impact as a compact scoreboard, from ken's own benchmark results in benchmarks/results/. One-shot display, not a persistent mode, and not a per-repo number. Trigger: /ken-gain, "ken gain", "what does ken save", "show ken impact", "ken scoreboard". |
Display the scoreboard when invoked. One-shot: do NOT change mode, write flag files, or persist anything.
The only figures this card may show are medians from ken's own runs recorded
in benchmarks/results/ of the ken repo. Render the recorded figures as plain
ASCII bars (bar length = measured range, label = exact figure).
If benchmarks/results/ has no ken results yet, print:
ken gain: no measured numbers yet.
Nothing is claimed that wasn't measured. Run the harness in benchmarks/
(see benchmarks/README.md; needs API keys in .env) to produce the first
real entry.
This repo, today: /ken-debt (the counted ceilings ledger)
/ken-audit (what's rewritable right now)
Figures are benchmark medians, not this repo. NEVER print a per-repo savings
number ("you saved X lines/tokens here"): the unbuilt version was never
written, so there is no real baseline to subtract from in a live repo. The
only real per-repo figures come from /ken-debt (a counted ledger), and
this card points there instead of inventing one. Never render another
project's numbers as ken's.
One-shot display. Edits nothing, changes no mode. "stop ken" or "normal mode": revert.