K
kaisa-kucherenko/claude-code-flow
My working Claude Code setup: output styles, hooks, skills, agents
无 SECURITY.md
在 GitHub 查看 ↗链接已于 2026-11-12 核查 · HTTP 200 · 打开的是厂商自己的官网
- ★ 数取值
- 25 ★ · kaisa-kucherenko/claude-code-flow · 取自 GitHub 搜索 API · (2026-11-13 20:00:50 核对)
- 来源明细
- 从该仓库的 GitHub 页面直接读取 · 条目类型:开源项目
- 收录与活跃
- 收录于 2026-10-21 · 最近见到 2026-11-14 · 共见到 28 次 · 被 1 个来源列出
- 评分构成
- 综合分 103 · 25 ★ · 1 个来源 · repo
常见问题(答案都来自我们的记录)
它适合谁 / 用来做什么?
我们把它归在这些标签下: 开源项目。标签来自条目自身,不是我们强加的分类。
谁维护它、还在更新吗?
仓库
kaisa-kucherenko/claude-code-flow;我们收录于 2026-10-21;最近见到 2026-11-14。这里没有"官方认证"或"质量评级"——我们只记录来源与时间。# claude-code-flow
以下摘录自该仓库的 README(片段) · 在 GitHub 读完整 README →
My working Claude Code setup — the output style, hooks, skills, and agents I
use daily as a tech lead doing AI-heavy development. Published as-is from my
own ~/.claude, sanitized of project specifics.
The problem all of this attacks: cognitive load. When AI writes the code,
you spend the whole day reading, verifying, and deciding — and the model's
default verbosity makes every one of those decisions more expensive. These
files are how I fight that.
Some files mix English instructions with Ukrainian examples — I work
bilingually. The mechanics are language-agnostic; adapt the examples to
yourself.
## What's here
| Component | What it does | Docs |
| Output style | BLUF, hard cut-pass, no closing summaries, colleague-not-assistant. Roy Kent from Ted Lasso as the persona. | output-styles/ |
| Persona re-inject hook | Fights persona drift on long contexts: re-injects a short style reminder into every prompt. | hooks/ |
| Development-cycle skills | adr → spec → implement → twix: decide with an explicit yes each, cut into phases, run one phase per commit, review with a neutral panel. | skills/ |
| Handoff / pickup skills | Deliberate session switching instead of /compact: you decide what survives, the next session verifies it against reality. | skills/ |
| Status line | The context-% number that drives the whole workflow, with traffic-light thresholds (green <30%, yellow 30–39%, red at 40+). | statusline/ |
| Agents | Implementers (Geralt/Lyutik), a three-model review panel (Agatha/Arthur/Dash), specialists (Jobs/Lauda/Gilfoyle), a diagram drawer (Ariel), a test pair (Sam writes, Dean breaks) with a shared pytest-quality skill. | agents/ |
| Precogs workflow | Orchestrates the panel: three reviewers in parallel, then a synthesis that dedupes and cross-votes findings. | workflows/ |
| Personas | Claude Code dressed as a character, one plugin each: a HUD above the prompt, sounds, themes, spinner lines, status line, and a /timer break reminder that opens a break screen. Two so far: AFC Richmond (Ted Lasso) and The Witcher 3. Linux, Ukrainian texts. | plugins/richmond/, plugins/witcher/ |
Each folder's README has the install steps and the reasoning; everything
targets ~/.claude. Run all install commands from the repo root. The personas
are a plugin marketplace instead:
claude plugin marketplace add kaisa-kucherenko/claude-code-flow
claude plugin install richmond@claude-code-flow
use daily as a tech lead doing AI-heavy development. Published as-is from my
own ~/.claude, sanitized of project specifics.
The problem all of this attacks: cognitive load. When AI writes the code,
you spend the whole day reading, verifying, and deciding — and the model's
default verbosity makes every one of those decisions more expensive. These
files are how I fight that.
Some files mix English instructions with Ukrainian examples — I work
bilingually. The mechanics are language-agnostic; adapt the examples to
yourself.
## What's here
| Component | What it does | Docs |
| Output style | BLUF, hard cut-pass, no closing summaries, colleague-not-assistant. Roy Kent from Ted Lasso as the persona. | output-styles/ |
| Persona re-inject hook | Fights persona drift on long contexts: re-injects a short style reminder into every prompt. | hooks/ |
| Development-cycle skills | adr → spec → implement → twix: decide with an explicit yes each, cut into phases, run one phase per commit, review with a neutral panel. | skills/ |
| Handoff / pickup skills | Deliberate session switching instead of /compact: you decide what survives, the next session verifies it against reality. | skills/ |
| Status line | The context-% number that drives the whole workflow, with traffic-light thresholds (green <30%, yellow 30–39%, red at 40+). | statusline/ |
| Agents | Implementers (Geralt/Lyutik), a three-model review panel (Agatha/Arthur/Dash), specialists (Jobs/Lauda/Gilfoyle), a diagram drawer (Ariel), a test pair (Sam writes, Dean breaks) with a shared pytest-quality skill. | agents/ |
| Precogs workflow | Orchestrates the panel: three reviewers in parallel, then a synthesis that dedupes and cross-votes findings. | workflows/ |
| Personas | Claude Code dressed as a character, one plugin each: a HUD above the prompt, sounds, themes, spinner lines, status line, and a /timer break reminder that opens a break screen. Two so far: AFC Richmond (Ted Lasso) and The Witcher 3. Linux, Ukrainian texts. | plugins/richmond/, plugins/witcher/ |
Each folder's README has the install steps and the reasoning; everything
targets ~/.claude. Run all install commands from the repo root. The personas
are a plugin marketplace instead:
claude plugin marketplace add kaisa-kucherenko/claude-code-flow
claude plugin install richmond@claude-code-flow
收录徽章(可选)
想把收录标记放进 README 就复制这行,可选,不影响收录。
[](https://anclum.com/zh/s/kaisa-kucherenko-claude-code-flow/)
- 来源
- github · https://github.com/kaisa-kucherenko/claude-code-flow
- 按领域浏览
- 开源项目
- Traction
- 25 ★
- 收录于
- 今天
- 安全策略
- 我们没在该仓库根目录找到
SECURITY.md(这不代表不安全)
同类推荐
所属领域与标签
每个标签都能点进去看同类的全部条目。数字是当前库里的真实条数。