workflow-skills

Builds a self-contained multi-source trending / hot-topic Markdown digest for any subject (AI, gaming, crypto, research, news, ...). Pulls fresh content in parallel from X/Twitter, Hacker News, Reddit, GitHub trending, arXiv, 微博 (Weibo), 36kr — plus any new source modules you drop into scripts/sources/. Use this skill whenever the user asks to find trending/hot/popular content on a topic, monitor what people are saying across social media or news sites, build a "top N posts" digest, aggregate AI/tech/research signals into a report, summarize cross-platform discussion, track Chinese (国内) + international (国外) coverage of an event, or curate a multimedia weekly digest with embedded images and videos. Triggers on phrases like "find hot posts about X", "X 热门帖", "trending now on AI", "monitor what people are saying about Z", "整理热点", "AI 资讯汇总", "Twitter + HN + Reddit trending report", "中文 + 英文 热点", "tech news roundup", "summarize discussion across platforms". The skill ships ready-made scrapers for 7 sources and is

Automates browser interactions for web testing, form filling, screenshots, and data extraction. Use when the user needs to navigate websites, interact with web pages, fill forms, take screenshots, or extract information from web pages.

Emit HTML reports (not Markdown) when the user asks for a structured report, analysis, plan, status, write-up, or dashboard. Loads a shared CSS shell from references/base-template.html and a 12-layer Claude-report skeleton from references/claude-report.html so output is body-only HTML (~40-55% token saving vs inline-CSS HTML). Use when the user requests: 报告/分析报告/调试报告/实现计划/状态报告/周报/日报/PR write-up/incident report/research explainer/dashboard; OR when output benefits from rich layout (cards, tables, stat bands, timeline, checklist, SVG, diff blocks). Do NOT trigger for: 简单回答/单段解释/纯代码片段/单 table 速查 — Markdown stays cheaper there.

Audit and improve CLAUDE.md files in repositories. Use when user asks to check, audit, update, improve, or fix CLAUDE.md files. Scans for all CLAUDE.md files, evaluates quality against templates, outputs quality report, then makes targeted updates. Also use when the user mentions "CLAUDE.md maintenance" or "project memory optimization".

Use when the user wants to BUILD a UI (page, component, dashboard, landing,

End-to-end methodology for distilling repetitive workflows into reusable automation. Mines conversation history and codebase for repeated patterns, implements project-specific automation, then extracts cross-project universal templates with sanitization and parameterization. Triggers when user mentions identifying repetitive operations, building reusable automation, extracting universal templates, finding what's worth automating, distilling workflows, asks "what can be automated" across a project, or wants to review/retire existing automation.

端到端自动化提炼方法论, 把重复流程蒸馏为可复用自动化。三阶段: 挖掘对话和代码库的重复模式 → 实施项目内自动化 → 提取跨项目通用模板 (含脱敏与参数化)。触发条件: 用户提到识别重复操作、构建可复用自动化、提取通用模板、找出可自动化的环节、流程蒸馏、问"项目里什么能自动化", 或想 review/退役已有自动化。中文版。

记录与追溯多版本数据/实验的派生血缘。用于回答"X 哪来的"、"X 排除了什么"、"X 和 parent 的差异是什么"。Triggers when user asks about data origin, derivation history, audit trail, removal/addition rationale, or completeness verification across versions.

从已有实验版本派生下一版, 保证单变量隔离 + 血缘记录。Triggers when user mentions deriving a new experiment version, bumping a config, single-variable ablation, or starting next iteration based on previous.

Walk a personal project through release end-to-end — scan changes, recommend a SemVer bump, preview every file write, wait for user confirm, execute, then ask separately before pushing. Use whenever the user says "release", "发版", "出版本", "tag a version", "cut v1.2.0", "let's ship", "make a release", "/release" — or implicitly when the user is finalizing CHANGELOG `[Unreleased]` and treating it as a version. Strongly prefer using this skill over ad-hoc tagging because all writes are previewed and reversible until the user types yes. Skip release-please workflow setup unless explicitly asked; this is the manual / interactive flow.

Use when the user wants to set up a fresh machine with their archived skills, restore skills after wiping ~/.claude/skills/, or sync newly-added skills from claude_workspace/workflow_skills/ to ~/.claude/skills/ in one shot. Triggering signals — phrases like "install all my skills", "bulk install skills", "set up skills on this machine", "restore my skills", or pointing at a workflow_skills/ archive and asking to install everything. Do NOT trigger for installing a single specific skill (the user can just cp that one folder).

当用户想在新机器上一次性还原所有归档 skill、清空 ~/.claude/skills/ 后批量恢复、或把 claude_workspace/workflow_skills/ 中新加的 skill 同步到 ~/.claude/skills/ 时使用。触发信号 —— "把我的 skill 全装上"、"一键安装所有 skill"、"在这台机器上配 skill"、"恢复 skill"，或者用户指向一个 workflow_skills/ 归档让你全部安装。**不要**用于只装单个 skill（用户直接 cp 那个目录更快）。

Build Python APIs with FastAPI, Pydantic v2, and SQLAlchemy 2.0. Covers project structure, async patterns, JWT authentication, validation, and database integration with uv package manager. Use when: creating Python APIs, setting up FastAPI projects, implementing JWT auth, configuring SQLAlchemy async, or troubleshooting 422 validation errors, CORS issues, or async blocking.

Use when given a multi-step research / coding / data-processing task that requires more than one round of implement → verify → fix, and you want closure without per-step human confirmation. Triggering signals — task says "run + analyze + report", "finish all of X", "iterate until done"; or you would otherwise pause to ask "is this what you wanted?" between every step.

收到需要多轮 实现→验证→修正 的研究/编码/数据处理任务、且希望 AI 自主闭环不在每一步等确认时使用。触发信号——任务里写"跑完 + 分析 + 出报告"、"把 X 全部做掉"、"迭代直到完成"；或者你不用这个 skill 时会在每个步骤之间停下来问"这是你要的吗"。

Use when bringing up an open-source GitHub research repo (any field) — cloning, configuring conda env, installing dependencies, downloading weights, running the demo. Triggering signals — README has a `quick_install.sh` / `requirements.txt` and you're about to follow it; pip install hangs / fails / pulls a different numpy unexpectedly; `import X` works but the demo crashes; demo "should take 30s" but takes hours.

配置任意开源 GitHub 研究仓库（任何领域）时使用——clone、配 conda env、装依赖、下权重、跑 demo。触发信号：README 给了 `quick_install.sh` / `requirements.txt` 你正要照着跑；pip install 卡住/失败/意外把 numpy 拉到不同版本；`import X` 能跑但 demo 崩；demo "应该 30 秒"却跑了几小时。

深度解析 AI 论文，生成可直接发布的专业阅读笔记

Elite 6-phase engineering production engine. Orchestrates specialized subagents (Think, Plan, Build, Review, Test, Ship) to deliver production-grade software with Garry Tan's "Boil the Lake" philosophy. Triggers on requests for feature engineering, project initialization, or full-lifecycle management. Uses structured gates for CEO/Staff/QA-level auditing.