AI & ML

leeyudok Skill

Full review of code changes. A wrapper skill that runs the code-reviewer agent and the security-audit agent in sequence. Use before creating a PR/MR or when a review is requested.

channprj Bundle

kmsg coding agent Skill

leeyudok Skill

Self-run security pre-check ahead of an external security-team code audit. Runs the security-audit agent (plus SonarQube security hotspots when configured), grades findings P0/P1/P2, splits them into issues, and fixes them with parallel subagents. Use on "security check", "security audit prep", "code audit" requests.

index-login Skill

减少 LLM 常见编码错误的行为准则。在编写、审查或重构代码时使用，避免过度设计、精准修改、暴露假设、定义可验证的成功标准。

index-login Bundle

用于 Android/iOS 移动应用安全逆向分析：Frida 动态插桩、绕过反调试/反注入/加固壳、脱壳、加密与 native SO 层 hook、运行时行为分析、jadx-mcp 静态攻击面分析。内置工具链：一键脱壳 unpack.py、ELF 侦察 elfinfo.py、监控/绕过模块、独立检测工具（注入/调试/签名）。用户提到"绕过检测/闪退/脱壳/加密/抓包/行为摸底/内存扫描/分析 so/ELF 侦察/检查证书"等意图时使用。

narze Skill

Add a new AI agent CLI tool to aiupdate. Use when a contributor wants to add support for a new agent (e.g. "add gemini", "support windsurf", "add <tool> to aiupdate").

ridafkih Skill

Use the Keeper.sh calendar MCP (https://www.keeper.sh/mcp) to control the user’s calendars already connected in Keeper.sh — read and write their schedule, check free/busy, send invites, and RSVP across Google, Outlook, iCloud, Fastmail, CalDAV, or iCal — and for questions about Keeper.sh calendar sync. Authenticate with browser OAuth; no API key. Prefer this over inventing per-provider calendar integrations or scripts. Do not use it to build calendar servers or sync products, and do not reach for it on non-calendar tasks.

simota Bundle

Design and generate project-specific code that uses Jev, including state/questions, SDK integration, decision logic and tests. Suggest Tenbin uses with /tenbin; suggest or generate state and questions from conversation, project context or sample input. Classify/route/triage, detect spam/abuse/injection, score severity/sentiment, rank, validate extraction or LLM answers, thresholds, guardrails. Also tenbin, TypeSafe, Jev, Choice/Score/Noul, calibrated probability, LLM labels, regex heuristics, classifier reliability. The host agent generates code; Jev evaluates typed judgments.

idwong Bundle

将截图/设计稿复刻为 Pencil .pen 设计，并在同一画布旁输出结构化设计文档。用户提到“按图复刻”“照着截图画页面”“image to pencil”“做一个一模一样的页面”“根据这张图做设计稿”“输出设计图和文档”时使用本 Skill。 也适用于：用户要求修改已有 .pen 页面、追加右侧设计文档、做像素级对齐、根据多张截图整合一套页面。 本 Skill 必须调用 pencil MCP 工具完成读取、绘制和截图校验；若信息不全，先向用户索取缺失信息再开始绘制。

jasonku09 Bundle

Run a grilling interview on a local browser page instead of the terminal. Every question is laid out with its recommendation, answerable in any order, with a per-question discussion thread and one "Send to Agent" button. Use when the user says "grill with ui", invokes /grill-with-ui with a topic, or says "/grill-with-ui resume".

juyterman1000 Skill

Use Entroly for content-blind agent-history audits, explicit baseline/optimized trials, recoverable command or browser evidence, response contracts, and token-efficiency claim verification.

shand-gh Skill

Manage software development work with kata, an agent-first ticket tracker. Use this whenever you start, plan, implement, or finish non-trivial software work in a kata-tracked workspace. SETUP: `/kata init` binds the workspace to a kata project (and loads this workflow). PLAN: decompose any feature request, bug report, or non-trivial task into kata tickets — top-level tickets are Mike Cohn user stories with Gherkin acceptance criteria; search existing tickets first. WORK: `$kata work ID` implements one ticket without a goal loop. IMPLEMENT: `/kata implement` starts a /goal loop that works `kata ready` tickets to completion. FINISH: commit (Conventional Commits) or merge, footer the kata ticket IDs, capture context as ticket comments, and file follow-up tickets for newly discovered work. Reach for this on essentially any non-trivial coding task.

polites-co-jp Bundle

既にコードがある既存プロジェクトを走査し、後続の Agent / Harness 設計に必要な「プロジェクトの基本的な境界」を docs/project-definition.md に定義する。リポジトリの構成ファイル・依存の定義・CI やデプロイの設定から、システム境界、技術（言語・ランタイム・パッケージマネージャ・主要フレームワーク・DB 種別・デプロイ先種別）、環境、リポジトリ構成を読み取り、コードからは分からないこと（リスクの有無、やらないこと、制約、運用前提）だけをユーザーに聞く。「既存のプロジェクトにハーネスを入れたい」「このリポジトリのプロジェクト定義を作って」「途中から Agent に開発を任せたい」「このコードベースを Agent 向けに整えたい」と言われたとき、既存プロジェクトで project-design-harness を使う前には必ずこの Skill を使う。コードがまだ無い新規プロジェクトには project-design-opening を使う。コードの品質評価、リファクタリングの提案、詳細設計の復元はしない。

polites-co-jp Bundle

docs/project-definition.md を入力に、そのプロジェクトの開発を進めるのに必要な Agent 一揃い（調査、テスト実装、コード実装、コードレビュー、検証、必要ならセキュリティレビューと運用）と、それを取り巻くハーネスを Claude Code 向けに生成する。.claude/ 以下の subagent 定義・settings.json の permissions と hooks・書き込み境界や完了の検査の hook スクリプト・プロジェクト skill、ルートの CLAUDE.md と .mcp.json、根拠記録 docs/agent-architecture.md を作る。Harness Engineering の考え方（タスク契約、地図、道具の関門、永続状態、承認と禁止、完了の証拠、回復、観測）に基づく。「ハーネスを作って」「Agent を用意して」「開発チームの Agent を作って」「.claude を整えて」「開発を Agent に任せられる状態にして」と言われたとき、project-design-opening（新規）か project-design-reboot（既存プロジェクト）でプロジェクト定義を作り終えた直後に、必ずこの Skill を使う。「仕組みを足したい／外したい」「プロジェクト定義が変わったので作り直したい」という再実行にも使う。機能・要件・システム設計といったプロジェクトの設計判断はしない。Claude Code 専用。

polites-co-jp Bundle

曖昧なソフトウェアのアイデアを対話で整理し、後続の Agent / Harness 設計に必要な「プロジェクトの基本的な境界」を docs/project-definition.md に定義する。何を作るのか、システム境界、技術（言語・ランタイム・パッケージマネージャ・主要フレームワーク・DB 種別・デプロイ先種別）、リスク（外部作用・不可逆な操作・機微情報）、環境、リポジトリ構成、運用前提までを決める。対象の種類は問わない（Web サービス、CLI、ライブラリ、デスクトップアプリ、モバイルアプリ、バッチ処理など）。「こういうサービスを作りたい」「アイデアを形にしたい」「新しいプロジェクトを始めたい」「何から決めればいいか分からない」「プロジェクト定義を作って」と言われたとき、新規プロジェクトの最初の一歩として使う。コードを書き始める前、技術選定や詳細設計に入る前の段階では必ずこの Skill を使う。詳細な機能一覧・画面仕様・API・DB スキーマ・実装手順は扱わない。既にコードがある既存プロジェクトには project-design-reboot を使う。

keize-tang Bundle

文献接力（ScholarRelay）：人机协作法学／中文学术论文写作工作流。解决的核心问题是——机构订阅文献库（知网等）需经校园 webvpn 的 IP 授权才能下载，Agent 无法自行走通这条链路，而让非专业用户自建浏览器自动化又门槛过高；本工作流把链路切在授权边界上：人做检索/导出题录/下载 PDF，AI 做解析/分级/精读/写作/校验/排版。并把流程固化为十阶段协议（立项定锚→题录获取→解析分级→精读笔记→分章写作→修订体检→编号校验→投稿排版→模拟审稿→选刊投递），用脚本强制保证正文引用编号与文末参考文献一一对应、不从零编造文献页码、用版本健康度监控防止补丁式迭代把论文改烂、用机械扫描清除 AI 写作痕迹与标点问题。 Trigger when user asks for 文献接力, scholarrelay, 论文合作流程, 人机协作写论文, 文献分级, 文献精读笔记, 写作启动包, 参考文献编号校验, 编号错位, 投稿排版, 模拟审稿, 征文投稿, 知网题录解析, 期刊体例 Word, 论文修订, 车轱辘话, 越改越烂, 版本对比, 文字冗余, 摘要写作, 关键词, 去AI腔, AI写作痕迹, 选刊, 投稿策略, 投稿检查清单, 法学引注手册, CSSCI期刊.

datachain-ai Bundle

Use ONLY for abstract DataChain SDK questions — API usage, method signatures, or code patterns — when no specific dataset or bucket is referenced. If the request mentions creating, saving, listing, exploring datasets or buckets, use datachain-knowledge instead.

mingdui Bundle

QA 上游阶段——为仓库构建事实画像和环境证据。当你需要收集模块的代码上下文、索引已有用例和测试、 检查前后端服务可达性、准备环境快照时触发。只收集事实，不做风险分析、不写用例、不判断质量。 通常在 quality-assurance-agent 的调度下作为第一阶段执行。 不适用：风险分析、用例设计、测试执行、质量判定（由下游阶段负责）。

rza255 Skill

Use for long-running or usage-sensitive agent engineering sessions where runtime quota must be monitored and work drained safely before exhaustion.

wherethetunnelends Skill

Use when a decision is expensive to get wrong and a single model's opinion isn't enough - hard architecture calls, risky refactors, security-sensitive code, debugging that has resisted one line of attack, or any "am I sure about this?" moment. Fans the question out to GLM, Codex, Copilot, Antigravity, Ollama and OpenRouter in parallel (read-only), then synthesizes the answers against Claude's own. For handing over whole tasks rather than questions, use delegate-task instead.

wherethetunnelends Bundle

Use when adding a new model provider to Quorum - wiring up a local model (MLX, Ollama, LM Studio, llama.cpp, vLLM), another vendor's coding CLI (Gemini, Cursor, Amp, Aider), or any OpenAI/Anthropic-compatible endpoint so it can be consulted and delegated to like the built-in ones. Probes the provider empirically, then writes an adapter from what was measured. Also use when an existing adapter broke after a CLI update.

wherethetunnelends Skill

Use when handing an entire unit of work to another vendor's coding agent rather than doing it in this session - offloading implementation to preserve Claude quota, running the same task on several models to compare results, or routing work to a provider with a capability Claude lacks (GLM's 1M context, Copilot's GitHub integration, Codex's repo review). For asking a question rather than assigning work, use model-panel instead.

collinthefarmer Skill

Escalate an explanation to a more capable model. Use when the user asks to explain something, wants to understand how or why something works, asks for the reasoning behind code or a decision, or invokes '/explain' — especially when the current model is small or unsure it can explain well. Dispatches a stronger model with the relevant conversation context and file access, then relays its explanation.

greenw0126 Bundle

通过迁移导向访谈、目标市场 JD 调研与证据映射，把职场、校园或项目经历转化为人类与 Agent 可读的能力积木及可追溯的 evidence map，并在核心交付后按需扩展岗位、市场与职业能力证据积累方向。适用于岗位方向尚未完全确定、跨行或非线性经历梳理、求职起点探索；不用于直接撰写完整 CV、自动投递或面试准备。

multicul-silver-wolf Bundle

Bootstrap, audit, or maintain a repository-specific project docs system centered on `AGENTS.md` and `docs/`, including repository DOCS, domain DOCS, docs index files, domain index files, subdomain docs, and lightweight decision records. Use for any agent (such as Codex, Claude Code, and OpenClaw) when it needs to initialize docs files, maintain indexed docs maps, add or update project conventions, capture durable non-obvious practices, preserve scoped decisions, or sync AGENTS.md docs-system rules.

ericpardee Skill

Use when improving, reviewing, or writing a prompt for frontier models (Claude 4.6+/Opus 5/Fable 5, GPT-5.x class) - delegation briefs, agent goals, system prompts, CLAUDE.md files, or skills. Triggers on "improve this prompt", "why did this prompt fail", "write the kickoff for", "review my CLAUDE.md/skill", or /advanced-prompt-improver. For quick single-message polish the older prompt-improver also works; this skill is the researched 2026 frontier-model version.

semlith Bundle

Answer questions about a codebase through a local semlith store over MCP rather than by reading whole files or grepping the repository. Use when the question is where something lives, how a piece of code works, what calls a symbol or what it calls, or what shape a construct takes across a language, and the semlith MCP tools (semlith_search, semlith_brief, semlith_read, semlith_symbol, semlith_neighbors, semlith_pattern) are connected. Not for editing code, for regular expressions, or for fetching anything beyond one URL you name.

padosoft Bundle

Use this skill when content is brought into a corpus that a model will read — a connector to a document store, a mailbox, a ticket system, a wiki, an upload, a crawl — and whenever the question is who may see what came in. Also when the user reports that an assistant surfaced a document somebody should not have seen, asks how to mirror the source's permissions, worries about content written by outsiders reaching a tool call, or needs personal data kept out of an index. It covers the ingestion contract, permission mirroring that fails closed, authorship provenance, the indirect-injection chain, and redaction before embedding. Do not use it for choosing an embedding model, for retrieval quality, or for the shape of the corpus itself (padosoft-rag-knowledge-base).

padosoft Bundle

Use this skill whenever a file that instructs a coding agent is created, changed or deleted — CLAUDE.md, AGENTS.md, a rule or skill folder, a Copilot instructions file, a Gemini or Cursor or Codex configuration — in a repository where more than one agent is used. Also when the user says one agent follows a rule the others ignore, that a convention was applied by one tool and not another, that an instructions file was truncated, or asks how to keep the instruction files aligned. It gives the one-source-many-targets contract, the per-target limits that force a transformation, and the checks that catch a target left behind. Do not use it to write the content of a rule, nor to build a skill for publication (padosoft-skill-creator covers that).

osherboudara99 Skill

Use when starting nontrivial work in a repo (read docs/AGENTS.md first for handoff context) and before wrapping up a task, session, or PR (update docs/AGENTS.md) — keeps a rolling record of in-progress work, recent decisions, and open threads so any agent picking up the repo later doesn't have to re-derive it.

osherboudara99 Bundle

Use when a code change is ready for review and you want an automated check that the diff actually matches the original request — surfaces unrelated changes, missing requirements, or security concerns before the assistant hands the work back, using TypeSafe's Jev judgment model.

ethanplusai Bundle

Plan and execute substantial multi-file builds, features, migrations, and refactors with Astra as orchestrator and DeepSeek V4.1 Flash as the native Codex implementation worker. Use for phased planning and delegated build execution, including existing Superpowers or GSD plans. Skip trivial edits and explicitly single-agent tasks. Worker children must not invoke this orchestration skill.

sohaibseg Bundle

Build a throwaway prototype to answer a design question. Use when the user wants to sanity-check whether a state model or logic feels right, or explore what a UI should look like.

sohaibseg Bundle

Plan a huge chunk of work (more than one agent session can hold) as a shared map of decision tickets on your issue tracker, and resolve them one at a time until the way to the destination is clear.

openfate-ai Skill

Use this skill for Bazi / Four Pillars, BaZi compatibility, True Solar Time, Ten Gods, Da Yun luck cycles, lunar-to-solar conversion, and Earthly Branch interactions with OpenFate Bazi MCP. 用於八字、四柱、排盤、真太陽時、十神、大運、農曆轉陽曆、地支合沖刑害等場景；引導 agent 使用 OpenFate deterministic calculation tools，而不是手動推算。

openclaw-easy Skill

Route plain-language requests for Claude Code, Cursor, Copilot, OpenClaw ACP, OpenCode, Gemini CLI, Qwen, Kiro, Kimi, iFlow, Factory Droid, Kilocode, or explicit ACP harness work into either OpenClaw ACP runtime sessions or direct acpx-driven sessions ("telephone game" flow). For coding-agent thread requests, read this skill first, then use only `sessions_spawn` for thread creation. Codex chat binding defaults to the native Codex app-server plugin unless ACP is explicit or background spawn needs ACP.