热门技能
当前社区里最受关注、增长最快的技能。
output-formats
kbrdn1
Skill migrated from _fragments/base/output-formats
refactor
luongnv89
基于 Martin Fowler 方法论的系统化代码重构 skill。适用于用户请求重构代码、改进代码结构、减少技术债、清理旧代码、消除 code smell 或提升可维护性时。这个 skill 采用分阶段、带研究与计划的安全增量实施方式。
diff-context
Barnett-Studios
"Use when the user asks to review changes, understand what changed, prepare a PR description, or needs context about recent modifications."
daily-update
prateek
Run the daily wiki maintenance cycle: check all source freshness, update the index, and regenerate hot.md. Use this skill when the user says "/daily-update", "run the daily update", "update everything", "morning sync", "refresh the wiki index", or when triggered by the launchd cron at 9 AM. Also use to set up or verify the cron + terminal notification infrastructure for the first time ("set up the daily cron", "install the terminal notification", "how do I get the morning reminder?").
me:design-pattern
kbrdn1
"Design patterns — select : cadrer un besoin, poser les questions qui tranchent, recommander le patron adapté (ou aucun). analyze : lire du code existant, nommer les patterns présents, les code smells et les violations SOLID, proposer des refactorings (read-only, aucune écriture). Garde-fou permanent : « pas de pattern » est toujours un candidat classé, pas une réserve de bas de page. Fondé sur GoF / refactoring.guru (23 patterns + 23 smells + techniques de refactoring), Fowler (Refactoring, PoEAA) et SOLID (R. C. Martin). Triggers: /me:design-pattern, design pattern, patron de conception, quel pattern utiliser, quelle architecture pour, SOLID, SRP, OCP, LSP, ISP, DIP, code smell, Strategy, Observer, Factory, Adapter, Decorator, State machine, over-engineering, sur-ingénierie, refactoring, dette technique."
agentcairn-setup
ccf
Install, configure, verify, upgrade, or repair AgentCairn persistent cross-agent memory. Use when a user wants shared local-first memory in Claude Code, Codex, Cursor, OpenCode, Antigravity, VS Code, Claude Desktop, or another supported harness; when AgentCairn recall or remember tools are missing; or when its plugin, MCP server, skill, or ambient hooks need setup diagnostics.
authentication-patterns
zebbern
"Authentication patterns: session vs JWT vs OAuth comparison, provider selection (NextAuth, Clerk, Supabase Auth), security checklist, and common mistakes. Use when implementing auth, reviewing auth flows, or choosing auth providers."
retrospective
taika-izumi
"サブプロジェクト1サイクル完了直後（feature ブランチを master へマージ後、handoff finalize 前）に実施する、サイクル末尾の課題抽出記録スキル。AI が plan / git log / handoff から課題候補（事象/原因/影響、system/flow 分類つき）を一括提示し、ユーザーが修正と起票判断を行う。フロー課題は delta 型（worklog で捕捉可能）か構造観察型かを振り分け、delta 型で急がないものは worklog パイプラインへ委ねる（ADR-0056）。抽出課題は docs/working/issues/system flow/ へ起票し（ADR-0028）、対策の採否・設計・ADR化は次サイクルへ委ねる（ADR-0021）。rubber-duck 独立レビューはユーザー要求時のみのオプション。"
writing-plugin-python
AtomiCloud
Write or modify CyanPrint plugin code in Python. Use when the user asks to add validation rules, change plugin behavior, modify the entry point, run commands from a plugin, or mutate files. Covers entry point (start_plugin_with_fn), native filesystem I/O (pathlib), and command execution (subprocess). Plugins receive { directory, config } and use native OS operations.
code-to-diagram
zebbern
"Analyze codebases and automatically generate architecture diagrams, flowcharts, and org charts. Uses AST parsing to map import dependencies for Python, JS/TS, Go, and Java, outputting Mermaid or SVG files. Triggered when users ask to visualize code architecture, understand dependencies, draw a flowchart, or create a module diagram from source code."
coding-harness-development
lmathia2
Implements and reviews Skein harness changes while preserving its one-tool notebook PTC, trace-native memory programs, replay, safety, and independently verified completion contracts. Use for orchestration, context, tools, repository access, durable state, verification, evaluation, and harness documentation.
ask-about-system
OpenSourceAGI
Guide to about-system (packages/about-system-info), the cross-platform system-info CLI and library — install and run it, pick specific info blocks, JSON output, the settings file (colors, emojis, labels, display order), the cache, the shell-greeting installer, and the programmatic API. Use when working with about-system or troubleshooting it — blocks that print empty, a greeting that doesn't run on terminal start, stale or wrong values, infoFunctions import errors, or --set changes that seem ignored.
code-critic
youlianvr
MUST USE when user asks to review code, find bugs, check for issues, audit code, code review, find bugs, check code, review, re-review, critique this, review this file, is this code good, what's wrong with, check for bugs, find problems. Sends code to proxy:4000 (deepseek-v4-flash) for merciless review. Returns structured report: score (0-10), critical count, each issue with severity/file/line/description/suggestion.
refactor
luongnv89
Систематичний рефакторинг коду на основі методології Мартіна Фаулера. Використовуйте, коли користувачі просять рефакторити код, покращити структуру коду, зменшити технічний борг, очистити застарілий код, усунути запахи коду (code smells) або покращити супровідність коду. Ця навичка проводить через поетапний підхід з дослідженням, плануванням та безпечною інкрементальною реалізацією.
me:loop:codex-review-pr
kbrdn1
Loop auto-cadencé — relance la review Codex CLI locale sur la PR de la branche courante et corrige jusqu'à 0 finding bloquant pertinent (P0/P1/P2), max 8 itérations, avec analyse de convergence pour s'arrêter quand les passes n'apportent plus rien. Déclencheurs : "/me:loop:codex-review-pr", "lance le loop codex-review-pr", "codex review jusqu'à ce que la PR soit clean".
debug-incident-protocol
youlianvr
Use when the user says: "doesn't work", "still broken", "hung/frozen", "disappeared after update", "used to work", "metric is zero but UI OK" — or when you need to investigate an incident, find the hang root cause, verify a process was actually restarted. Covers: facts before theories (storage/logs/PID), symptom vs root cause, silent failure by metrics, single consumer, restart ritual, hang localization via progress marker, timeouts, cache masking. Do not use for test writing (testing-discipline) and refactoring (agent-refactor-safety).
assets
arabold
Use when discovering, downloading, or adding new 3D/2D/audio assets to this game from curated free sources (KayKit, Quaternius, Kenney, Poly Pizza) or generating our own via image-to-3D. Covers the two automation channels (KayKit GitHub + Poly Pizza API), the download->manifest->import workflow, licensing/style rules, and the experimental AI authoring path. Policy and the asset manifest live in assets/AGENTS.md; this skill is the how-to and the scripts. Not needed for editing assets already in the repo.
dataviz
youlianvr
Choose and produce an appropriate chart/dashboard/visual explanation from data, with legible encodings and source/assumption notes.
speckit.plan
kbrdn1
Execute the implementation planning workflow using the plan template to generate design artifacts.
extend-guidelines
taika-izumi
"ガイドラインの拡張（原則追加・Skillの新規作成/改定・AGENTS.md更新）を行う際のゲートウェイ。CONTRIBUTING.mdを読み込み、brainstormingへ接続する。"
writing-processor-javascript
AtomiCloud
Write or modify CyanPrint processor code in JavaScript. Use when the user asks to change file transformations, modify the entry point, handle file processing, or change output generation. Covers entry point (StartProcessorWithLambda), CyanFileHelper (read/write/copy/resolveAll), and VirtualFile (content/read/write). Processor lambda receives (input, fileHelper) as two parameters.
pstack
Sma1lboy
Rigorous engineering mode for nontrivial work in this repo — a set of named principles plus the leaf skills that apply them. Use when the user says "pstack", "go deep", "be rigorous", "认真做", or when a task involves architecture, a real bug, a refactor, or anything the user will not be watching. Ported from cursor/plugins pstack by Lauren Tan (MIT); see ATTRIBUTION.md.
dolphin-windows-theme
youlianvr
Make the Dolphin file manager (flatpak on GNOME/Fedora) look like Windows 11 File Explorer — dark gray #1F1F1F background, white text, blue #0078D4 selection. Use when the user says "dolphin dark theme", "dolphin like on Windows", "like Windows Explorer", "white text dolphin", "dolphin not dark", "theme dolphin dark", "windows 11 dark dolphin", or a flatpak Qt/KDE app ignores ~/.config/kdeglobals and stays light. Covers kdeglobals/color-scheme placement inside the flatpak sandbox, QT_QPA_PLATFORMTHEME=kde, flatpak overrides, and pixel-verification without seeing the screen.
me:loop:test-until-green
kbrdn1
Loop auto-cadencé générique (tout langage) — détecte le runner de tests du projet, le relance et corrige jusqu'à ce que tous les tests passent (exit 0), max 10 itérations. Déclencheurs : "/me:loop:test-until-green", "lance le loop test-until-green", "boucle jusqu'à ce que les tests soient verts".