agentleFS
kizemokizemo/evidence-skill/SKILL.md

evidence-skill

Skill1 starsChanged 18 days agoMITView on GitHub

Use when configuring, debugging, or designing Evidence BI Core 2026 pages on bi.jury.xin — charts, tables, filters, Chinese-schema SQL, nginx, themes, MCP, upgrades, and report-page UI contracts.

1 other project uses a version of this.

Adopted by

Most recently taken up by

  • Other versions1 project uses a version of this.

    Word for word · 0

    None found.

    Edited · 0

    None found.

    Wrote their own version · 1

    1. awslabs/agentic-developer-platform★ 7same name

    How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.

The file

Files
SKILL.md

Evidence Core 2026 — Sales BI

改完必验收 3-piece(SQL 直查 + Playwright 真渲染 + 像素采样)——「配置对了」不等于「渲染对了」。

Source: mirror/(docs.evidence.dev 公开文档镜像,137 页)+ templates/ + GBrain;契约以 prod 二进制实证为准,不采信文档推测。作者本机原始 vault = E:/obsidian/evidence/。 Generated: 2026-10-13 初始 → 2026-10-14 fold-in(R1-R21)→ 2026-10-21 v0.3(补 ch21 报表页 UI 契约)→ 2026-10-25 v0.3.1(token 瘦身)→ v0.3.2(公开文档镜像 mirror/ 随 skill 分发,远程安装自包含)

When to use / NOT

用本 skill不用本 skill
Evidence Core 0.10.1 页面/图表/表格/SQL/部署/升级/主题/MCP/报表页 UI 契约Legacy 40.x 新功能 · Studio 专属 · 非 PG · 纯 CSS 调优(用 impeccable)

Chapter Index

21 章按需加载;主题 → 章的完整映射见 references/topic-index.md。

章组覆盖(章号 → 主题)何时读
ch01-ch10ch01 架构 · ch02 config · ch03 监控 · ch04 cutover 失败 · ch05 SQL 反模式 · ch06 nginx patch · ch07 postgres_fdw · ch08 样式 · ch09 复盘 · ch10 运维速查动 prod 前
ch11-ch18ch11 组件 · ch12 部署 · ch13 鉴权 · ch14 性能 · ch15 主题真相 · ch16 FastMCP · ch17 升级 SOP · ch18 BI 工具选型系统性设计
ch19-ch20ch19 94 组件全量索引 · ch20 Evidence Warehouse 单页真相找组件 / 架构
ch21报表页 UI 契约(图表 14 条 · 表格 · 页面内容 · 3-piece · A/B 变体法)改报表页

References(按需加载)

文件何时读
references/cheatsheet.md16 决策表,决策快查(最高优先级)
references/patterns.md8 大模式 + 反模式表,复用现成方案
references/glossary.md术语不熟
references/frameworks.md需要解释"为什么"
references/topic-index.md80+ 主题 → 章映射,grep 不到时
references/protocol.md交付答案格式 · 出错先查
mirror/94 组件原文 + 8 类公开文档镜像(随 skill 分发,离线可查)
templates/踩坑/调研/决策 3 个写回模板(复制到你的 vault _playbook/)

Scope & Limits

覆盖: Core 0.10.1 配置/部署/排错/升级/性能/主题/MCP/报表页 UI 契约;中文 PG schema;nginx patch;vs Metabase/Lightdash。 不覆盖: Legacy 40.x 新功能;Studio 专属;非 PG;多 Core;Pentagram 深度设计。 已知限制: cryptography 包损坏 (Python 3.14);13+ page 待修列名 bug;gbrain keyless 无 semantic search。 升级频率: 每周 changelog BREAKING;每月 PG slow query;每季 Core 升级 (14 天 hold-back)。

Anti-Patterns(踩坑飞轮)

真实踩坑 = 已验证 ≥1 次;不收未验证断言。

看到别做真做
中文 "店铺" 直接 SQL不加双引号必须 "店铺" 包裹 — ch05
nginx proxy_pass 改完1 行 sed 重启5 处 patch + 5 秒回滚 — ch06
改 prod 页面直接 scp备份 + diff;备份必须对应要回到的状态 — ch21 §4
图表改完看配置就说好了3-piece 验收 — ch21 §5
想写「怎么改报表」的说明写进报表页放 json / memory / handoff — ch21 §3
同步脚本加正则用位置组具名组 + 注入漂移往返测试 — ch21 §5
agent 查 Evidence暴露 SQL 不拦截FastMCP 4 道防线 — ch16
sidebar 找不到组件推断 sidebar=sitemapBFS+sitemap 混合 — ch20

Verification

  • wc -w SKILL.md ≤ 1000(重内容在 chapters/ 与 references/)
  • frontmatter 可被 YAML 解析(description 用块标量)
  • 新契约落在 chapter,没堆在 SKILL.md
  • 改 prod 前后都跑 3-piece
  • 沉淀已回写 Obsidian / memory

沉淀回写 SOP(答完/踩坑后主动做)

触发立即做模板
切 prod / 升级失败 / ≥2 次重试未修写 <vault>/YYYY-MM-DD-bug-<short>.mdtemplates/踩坑模板.md
新技术 / 新工具 / 重大变更前调研写 <vault>/_research/R<NN>-<topic>.mdtemplates/调研模板.md
不可逆 / ≥1 月影响决策写 <vault>/_decisions/YYYY-MM-DD-<short>.mdtemplates/决策模板.md
组件契约新发现补 chapters/ch21-*.md(图表/表格契约)契约表 + 踩坑表
完成后(可选)gbrain remember fact="..." entity=projects/sales-bi—

模板位置:templates/(随 skill 分发;复制到你的 vault _playbook/ 即可用,模板里的 ../_research/ 是相对你的 vault 的路径)。作者本机 vault = E:/obsidian/evidence/。 门槛(满足 ≥2 条才写):不可逆 / 影响 ≥1 月 / 多个合理方案 / 未来 agent 重做必须知道原因。 周回写:每周五或下次 session 开头自审 3 问 —— 这周有什么值得沉淀?属踩坑/调研/决策哪类?有 → 填模板。

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.