CAT
/Skills
SkillsMCPMarketplacesDigestToolsAdvertise

This week in Claude

Every Monday: Claude Code, Agent SDK, MCP, and the Anthropic platform moves worth your time.

Skills by Category
Frontend DevelopmentBackend & APIsTesting & QASecurityDevOps & CI/CDGit & Pull RequestsDocumentationCode Review & QualityAI & Agent BuildingSkill Development
MCP Servers by Category
Sales & MarketingWeb & Browser AutomationDatabasesAI & LLM ToolsCloud & InfrastructureCommunication & MessagingDeveloper ToolsDesign & CreativeDocuments & KnowledgeSearch & Web Crawling
Marketplaces by Category
AI Agents & OrchestrationLLM IntegrationDevelopment ToolsFrontend & UIBackend & APIsDatabasesTesting & Code QualityDevOps & CloudSecurity & ComplianceGit & Version Control

Cross AI Tools

Discover Claude Code plugins, extensions, and tools. Automatically updated directory of Anthropic Claude AI marketplaces with development tools, productivity plugins, and integrations.

Resources

  • Browse Skills
  • Browse MCP Servers
  • Browse Marketplaces
  • Plugins Reference

Community

  • About
  • Tools
  • Feedback
  • Privacy Policy
  • Advertise

Built for the Claude Code community with Claude Code by @mertduzgun

Independent project, not affiliated with Anthropic

Cs Issue Report

liuzhengdongfortest/codestable
905 installs890 stars
Summary

This is the first stage of a structured issue workflow that turns a user's vague "I found a bug" into a properly documented report. It walks through five mandatory questions,what happened, how to reproduce it, expected vs actual behavior, environment info, and severity,while enforcing a strict rule: capture only observable phenomena, never root cause speculation. That distinction matters because mixing user hunches into the report sends stage 2 analysis down the wrong path. The skill also gates whether an issue takes the fast track (obvious one-liner fix) or standard path (needs deeper investigation). It's opinionated about forcing concrete answers instead of accepting "sometimes it breaks" and refuses to let you skip ahead into diagnosis before the report is locked in.

Install to Claude Code

npx -y skills add liuzhengdongfortest/codestable --skill cs-issue-report --agent claude-code

Installs into .claude/skills of the current project.

CodeRabbit
CodeRabbit
AI writes the code. CodeRabbit catches the slop.
Try For Free →
Keep your Mac awake
Keep your Mac awake
Keep your Mac awake while Claude Code and 40+ AI agents run. Sleeps when they're idle.
One time payment $9 →
Context.devContext.dev
Context.dev
Integrate web data into your AI product. One API to scrape website & brand data.
Get API Key Now →
Make your agent a DeFi expert
Make your agent a DeFi expert
Agent, run crypto. Access onchain data & trade routes via 1inch.
Install now →
Make money from your Skills
Make money from your Skills
On Capafy, your Skill runs online 24/7 as an agent product, and you get paid every time someone uses it.
Start earning →
AppSignal
AppSignal
Monitor with ease. Code with confidence.
Start Free Trial →
CodeRabbit
CodeRabbit
AI writes the code. CodeRabbit catches the slop.
Try For Free →
Keep your Mac awake
Keep your Mac awake
Keep your Mac awake while Claude Code and 40+ AI agents run. Sleeps when they're idle.
One time payment $9 →
Context.devContext.dev
Context.dev
Integrate web data into your AI product. One API to scrape website & brand data.
Get API Key Now →
Make your agent a DeFi expert
Make your agent a DeFi expert
Agent, run crypto. Access onchain data & trade routes via 1inch.
Install now →
Make money from your Skills
Make money from your Skills
On Capafy, your Skill runs online 24/7 as an agent product, and you get paid every time someone uses it.
Start earning →
AppSignal
AppSignal
Monitor with ease. Code with confidence.
Start Free Trial →
Files
SKILL.mdView on GitHub

cs-issue-report

启动必读

开始任何判断或动作前,先读取 .codestable/attention.md;缺失则视为骨架不完整,提示先补齐或运行 cs-onboard,不要回退到外部 AI 入口文件。

这一阶段做两件事:把用户脑子里的问题落成结构化记录 + 判断走标准路径还是快速通道。

核心原则:只记现象不记根因。用户说"我觉得是 XX 组件的问题"——记下"用户怀疑 XX 组件"作为线索,但不顺着聊根因。根因要在阶段 2 通过实际读代码确认,不靠脑子里猜。混进根因猜测的报告会带偏阶段 2,让分析人围着错误线索绕。

共享路径与命名约定看 .codestable/reference/shared-conventions.md 第 0 节和 cs-issue 的"文件放哪儿"。


启动检查

  1. 确认是 bug 不是新功能需求——描述"想加 X 功能"的告诉他走 cs-feat
  2. 看有没有相关 issue 目录——Glob .codestable/issues/,有同类问题先和用户确认是新建还是更新
  3. 快速通道判断(唯一正式判定点)——按用户线索读一下相关代码(Grep / Read 定位):
    • 能一眼确定根因(能给出 {文件}:{行号}、修复改动小 1-2 处、无跨模块影响风险)→ 告诉用户"我已看到问题所在,可以走快速通道:直接告知根因和修复方案,你确认后我立刻修,修完你验证,只写一份 {slug}-fix-note.md"。同意后触发 cs-issue-fix(快速通道模式)
    • 不能(根因有多个候选 / 不确定 / 需要更多复现信息)→ 走标准路径做完整问题报告。进入标准路径后默认不再二次改判
  4. 确定 issue 目录名——跟用户商定 slug,日期前缀用今天(环境信息 currentDate)。目录不存在就创建。快速通道也要建 issue 目录,{slug}-fix-note.md 放那

必答 5 问

按顺序逐个问,不一次抛 5 个——一次抛多个用户只回最容易的,深的就漏了。每问做一次模糊度检查,不通过继续追问。

1. 问题是什么?看到的现象?

期待具体的异常表现:"点击提交按钮后弹出了空白弹窗" 比 "提交功能有问题" 有用一百倍。

模糊信号:"有时候会出错"、"感觉不对" → 追问"具体什么时候"、"具体什么不对"。

红线:不要让用户描述根因。出现"应该是因为 XXX"——记现象,根因留给阶段 2。

2. 怎么复现?

期待最小复现步骤:进入 XX 页面 → 输入 YY → 点击 ZZ → 看到问题现象。

模糊信号:"不稳定复现"、"有时候能有时候不能" → 追问复现频率和条件差异。确实不稳定就写明已知触发条件和复现率。

"无法复现"也是有效回答——写"目前无法稳定复现,只在 X 情况下观察到一次",别勉强凑步骤。

3. 期望行为 vs 实际行为

期待两句话:

  • 期望:我以为做了 A 之后应该发生 B
  • 实际:但实际发生了 C

不要合并成一句。合在一句"按钮没正常工作"里分析的人不知道"正常"长什么样。

4. 环境信息

最低采集:在哪个模块 / 功能区域发现的、相关文件或函数(用户知道的话)。

可选:操作系统、浏览器版本、运行环境(dev / prod)、最近有没有改过相关代码。

用户说"不知道在哪个文件"——写"待定",阶段 2 分析时查。

5. 严重程度与优先级

  • P0 阻塞:核心功能完全失效,影响所有用户,必须立刻修
  • P1 严重:核心功能受损有绕过方法,尽快修
  • P2 中等:非核心功能或影响少数用户,计划内修
  • P3 轻微:UI 瑕疵 / 边界情况 / 有更好实现,按空闲修

用户不确定就帮他推荐一个但让用户拍板。


问题报告模板

---
doc_type: issue-report
issue: {issue 目录名}
status: draft
severity: P0 | P1 | P2 | P3
summary: {问题现象一句话}
tags: []
---

# {问题简述} Issue Report

## 1. 问题现象

{用户描述的具体异常表现,纯现象描述,不含根因推测}

## 2. 复现步骤

1. {步骤 1}
2. {步骤 2}
3. 观察到:{问题现象}

复现频率:{稳定 / 概率(约 X%) / 暂无法稳定}

## 3. 期望 vs 实际

**期望行为**:{做了 A 应该发生 B}

**实际行为**:{但实际发生了 C}

## 4. 环境信息

- 涉及模块 / 功能:{模块名或功能描述}
- 相关文件 / 函数:{已知 file:line 或"待定"}
- 运行环境:{dev / staging / prod / 不确定}
- 其他上下文:{OS、浏览器、最近改动等,没有写"无"}

## 5. 严重程度

**{P0 / P1 / P2 / P3}** — {一句话理由}

## 备注

{可选:截图描述、日志片段等}

退出条件

  • frontmatter 完整(doc_type=issue-report / issue 一致 / severity 和 summary 非空 / tags ≥ 1)
  • 5 问都有具体答案(环境中相关文件可"待定")
  • 问题现象是纯现象描述,没混入根因推测
  • 期望 vs 实际显式分开写
  • 复现步骤可执行("无法稳定复现"也有说明)
  • 用户明确说"report 可以了,进下一步"
  • frontmatter status: confirmed

退出后

告诉用户:"问题报告已就绪。下一步阶段 2 根因分析,触发 cs-issue-analyze。"

别自己顺手开始分析根因——阶段间的人工 checkpoint 是工作流硬约束。


容易踩的坑

  • 用户说"可能是 XX 的问题"你顺着聊根因——错,那是阶段 2
  • 复现步骤太模糊("在用户界面操作一下")就放行——逼出可执行步骤
  • 期望和实际混在一段话里——必须显式分开
  • 严重程度留空——给默认值或写"无"
  • 一口气把 5 问列成清单丢给用户填——逐题对话否则深的全漏
Featured
CodeRabbit
CodeRabbit
AI writes the code. CodeRabbit catches the slop.
Try For Free →
Keep your Mac awake
Keep your Mac awake
Keep your Mac awake while Claude Code and 40+ AI agents run. Sleeps when they're idle.
One time payment $9 →
Context.devContext.dev
Context.dev
Integrate web data into your AI product. One API to scrape website & brand data.
Get API Key Now →
Make your agent a DeFi expert
Make your agent a DeFi expert
Agent, run crypto. Access onchain data & trade routes via 1inch.
Install now →
Make money from your Skills
Make money from your Skills
On Capafy, your Skill runs online 24/7 as an agent product, and you get paid every time someone uses it.
Start earning →
AppSignal
AppSignal
Monitor with ease. Code with confidence.
Start Free Trial →
Categories
Git & Pull Requests
First SeenMay 16, 2026
View on GitHub

Recommended

More Git & Pull Requests →
github-pr-review

fvadicamo/dev-agent-skills

github pr review
491
63
github-pr-merge

fvadicamo/dev-agent-skills

github pr merge
214
63
make-pr-easy-to-review

cursor/plugins

Prepare PRs for review by cleaning noisy history, improving PR descriptions, and adding reviewer guidance without changing code behavior.
2.1k
git-commit

github/awesome-copilot

Standardized git commits using Conventional Commits specification with intelligent diff analysis and message generation.
33.7k
34.3k
pr-review-expert

alirezarezvani/claude-skills

Performs systematic code review of GitHub PRs and GitLab MRs with blast radius analysis, security scanning, breaking change detection,...
534
16.9k
pr-review

microsoft/win-dev-skills

Multi-dimensional review of a PR or feature branch in microsoft/win-dev-skills. Activate on "review my PR / changes / branch", "vet before pushing", "PR review", "is this ready to merge". Fans out parallel sub-agents over skill content, the skill-vs-tool boundary (solution hierarchy), tool correctness, payload/provenance/tests, docs & manifest sync, plus a multi-model cross-check. Reports findings to stdout. Does NOT apply fixes.
288