Skip to content
Agentmart
Submit a tool

SK-model-maintenance

模型维护最小化纪律(用户提议·借鉴 Ponytail 的 YAGNI 内核)——**仅用于「更新模型」场景**(改脚本/加功能/写文档/重构/加数据源),**不用于比赛分析**。核心:最小实现·能不加就不加·能复用就复用·用完即清;红线:三处文档同步/check_sync/备份…

Skill by MENG-COOLMAN

IndexedMIT★ 3
Write a review

Chalk Labs

Need help deploying SK-model-maintenance?

Chalk Labs sets up agents like this inside real teams: connected to your tools, tested on your data, with a human in the loop where it matters.

Book a call

Install

Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.

npx agentmarket install skill/meng-coolman-sk-model-maintenance

Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.

Get the weekly diff

The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.

We use your email only for this newsletter. Privacy

About

模型维护最小化纪律(用户提议·借鉴 Ponytail 的 YAGNI 内核)——**仅用于「更新模型」场景**(改脚本/加功能/写文档/重构/加数据源),**不用于比赛分析**。核心:最小实现·能不加就不加·能复用就复用·用完即清;红线:三处文档同步/check_sync/备份 不可"懒"。用户说"更新/改/加/精简模型"时读本 skill;**分析场景不读**(用户发 txt 分析时)。

No reviews yet. Installed it? Reviews come from verified installers only. How reviews work

Permissions

What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.

Tools
No tools found
Secrets
None asked for
Install scripts
None found

Network hosts it talks to 1

  • codeload.github.com · found in install command

Files it touches

  • ~/.claude/skills/sk-model-maintenance

What's inside

Skill name
SK-model-maintenance
Description
模型维护最小化纪律(用户提议·借鉴 Ponytail 的 YAGNI 内核)——**仅用于「更新模型」场景**(改脚本/加功能/写文档/重构/加数据源),**不用于比赛分析**。核心:最小实现·能不加就不加·能复用就复用·用完即清;红线:三处文档同步/check_sync/备份 不可"懒"。用户说"更新/改/加/精简模型"时读本 skill;**分析场景不读**(用户发 txt 分析时)。
License
MIT
Instructions
58 lines in SKILL.md · view on GitHub

Agent Skills spec issues

  • name: only lowercase letters, digits and single hyphens allowed

Files (1)

  • SKILL.md

SKILL.md (first lines)

# SK-model-maintenance — 模型维护最小化纪律

## 🔴 适用范围(场景隔离·唯一触发点)

- ✅ **适用**:**更新模型时**——改脚本 / 新增功能 / 写文档 / 重构 / 加数据源 / 清理归档
- ❌ **不适用**:**比赛分析(Step0-11)**——分析场景由「预算红线豁免铁律」管辖:
 **零步骤省略·零输出弱化·token/篇幅永远不是理由·以正确性为目的**
- 🔴 触发词区分:用户说「更新/改/加/精简/清理模型」→ 读本 skill;用户发竞彩 txt 要分析 → **不读**(按原流程完整执行)

## 🔴 YAGNI 阶梯(动手前逐级自问·借鉴 Ponytail)

```
① 这真的需要存在吗?        → 不需要 → 跳过(多数需求止步于此)
② 已有脚本/函数/表能复用吗? → 复用(先查: data/tmp 脚本 · data/online_learning 模块 · 现成 json 表)
③ 标准库/现成参数/机制能解决吗?→ 用(新增 argparse 参数 ≫ 新写脚本)
④ 最小实现                   → 够用即止·不预留"以后可能用得上"的扩展/抽象层
⑤ 写后自查                   → 删除零消费的代码/参数(存在≠活跃·判定标准=产物是否被消费)
```

## 🔴 不可"懒"的红线(维护场景同样强制·违反即未完成)

| 红线 | 说明 |
|:--|:--|
| **文档三处同步** | AGENTS + 主技能 SKILL.md + `data/足球分析模型.txt` ——缺一不可 |
| **check_sync 全绿** | 含新增防回退校验(改了什么就加什么校验) |
| **改前备份** | `data/backups/<时间戳>/`(含被改文件) |
| **副本同步** | 涉及桌面副本已同步的文件 → 同步 + 双端 check_sync |
| **用完即清** | 临时脚本/探针/测试文件 → 删除(不留仓库) |
| **数据源纪律** | 按「数据源成功路径固化表」执行(如 api-football 纯 date) |

## 🔴 数据链三层风险(用户核心关切「一切以模型数据链不出错为基准」)

**任何改动先问:动的是哪一层?**

| 层 | 内容 | 改动纪律 |
|:--|:--|:--|
| **① 输入/解析层** | txt 竞彩格式解析(`txt_normalize`/`txt_features`)、欧亚盘抓取(Odds-API/api-football) | 🔴 **高危·禁止擅动**——字段错位/静默失效/数据提不出(真实事故: goal_odds "球后缀"静默失败 / 让球标题正则不匹配致假"数据不足" / F8 格式不匹配恒失效)。改前必须**用真实 txt 双格式实测** |
| **② 数据表层** | `goal_bins_table.json` / `league_table.json` / `score_depth_table.json` 等(键名+结构) | 🔴 **高危·禁改键名**——下游 live 引擎/查表按列名读 → 改键=静默失效。新增只能**加值不改名**·并同步全部消费方 |
| **③ 展示/留痕层** | calc_all 输出行 / raw 第八节 / 模板 / 校验器(output_checker/check_luopan) | 🟢 **安全区·新增标注优先放这层**——只读、不改数据、不改接口 |

**铁律**:**新增任何「标注/别名/档位显示」= 只读消费现有表 → 一律放第③层**;**外部术语体系(如他处提到的 L1-L7 七档量级)不得反向改写第①/②层**——量级与口径以 **txt 竞彩(总进球 8档)+ 欧亚盘(O2.5)为唯一基准**(决策: L1-L7 不引入·原表不在库不造数字 + 三套档位并存=口径混乱风险)。
**自检**:改完后问「这条改动会不会改变任何**已有字段的读法**?」→ 会 → 停手重设计。

## 🔴 实证(为什么需要本纪律)

- **2026-09-19 一轮内**即产生 2 个临时运行器(`_prestep_run_0917.py` / `_ps917_night.py` → 用后已删)——正是本纪律要防的"多写"
- **历史归档**:8 个脚本因"产物零消费"归档 → 本纪律的第 ⑤ 条(写后自查消费方)可提前拦住
- **对照正例**:`txt_normalize.py`(330 行)= **真正需要的工具**(下游实测揪出 2 个真 bug·格式兼容覆盖 6 类输入)→ **合理规模·非过度工程**
- 🔴 **注意区分**:分析场景的「批末压缩」「分 5 批执行」是**上下文管理**(信息已持久化 raw·零损失)·**不属于**本纪律的"精简"范畴

## 📋 维护任务标准流程(最小版)

```
① 备份 → ② YAGNI 阶梯自问(能否复用/最小实现) → ③ 修改(代码+文档三处)
→ ④ check_sync(+防回退校验) → ⑤ 副本同步+双端全绿 → ⑥ 清理临时产物 → ⑦ 记忆(一条)
```

From MENG-COOLMAN/PitchQuant, MIT.

More from MENG-COOLMAN/PitchQuant

Signals

Stars
3
Forks
2
Open issues
0
Last commit (this folder)
Sep 19, 2026 (1 month ago)

From GitHub (fetched Oct 24, 2026).

Details

License
MIT
First seen
Oct 24, 2026
Last checked
Oct 24, 2026

More from MENG-COOLMAN

Done for you

Rather not set it up yourself?

Chalk Labs scopes, builds and runs agents for your team: the task audit, the eval harness, the integrations and the weekly ops report. Fixed scope, priced before we start.

See how we build AI agents →Read the case studies

Chalk Labs is a two-founder AI and growth studio. We build AI agents, CRMs and MVPs for Web3 and AI companies, and have shipped 50+ product and token launches.

launches shipped
50+
launches shipped
media placements
500+
media placements
founders on every account
2
founders on every account

Press we've placed: Forbes · CoinDesk · Decrypt · Cointelegraph · The Block

Rahil Jain, Co-founder, Chalk Labs. Ex-Goldman Sachs engineer. You talk to the person who builds it.

Get SK-model-maintenance running in your business

Tell us what you want to automate. A founder replies within one working day with a plan, not a pitch deck.

Goes to the Chalk Labs founders, not the newsletter. Privacy

Need help deploying SK-model-maintenance?

CallGet help
Cite this page

SK-model-maintenance skill. Agent Mart. https://agentmart.dev/skills/meng-coolman-sk-model-maintenance (last updated 2026-10-24)