Context & AI engineering
25,831 rows, by stacks then stars
1,865 of these skill files have been read, by 1,328 distinct authors, and what they tell an agent counted. What Context & AI engineering authors agree on, and what they forbid
magnus919/agent-skills/agent-evals-and-observability Skill
169★ repoDesign, run, review, or release framework- and vendor-neutral evaluations and observability for AI agents. Use when defining agent evals, datasets, graders, trajectory review, regression analysis, release gates, production traces, or privacy-aware telemetry. Covers task and trajectory contracts, statistical comparisons, and incident-to-case learning; route framework implementation to pydanticai or langgraph when needed.
magnus919/agent-skills/agent-skills Skill
169★ repoUse this skill when creating, reviewing, or editing Agent Skills-format skills, or when implementing skill discovery and loading in an agent client. It covers directory structure, SKILL.md metadata, progressive disclosure, evals, and repository conventions. Do not use this skill for general software work that does not involve the Agent Skills format or lifecycle.
magnus919/agent-skills/crewai Skill
169★ repoExpert skill for role-based multi-agent orchestration with CrewAI. Agents with Role/Goal/Backstory, task design, crew composition (sequential or hierarchical), tool integration, callbacks, and production deployment. Use when orchestrating multi-agent teams or comparing agent frameworks.
magnus919/agent-skills/dspy Skill
169★ repoExpert skill for programming—not prompting—language models with Stanford's DSPy framework. Signatures, modules (Predict, ChainOfThought, ReAct), optimizer/teleprompter selection, compilation, caching, evaluation. Use when doing programmatic prompt optimization or building compiled prompt programs.
magnus919/agent-skills/langchain Skill
169★ repoExpert skill for building LLM applications with LangChain — LCEL chains, RAG pipelines, agent orchestration, LangGraph integration, LangSmith observability, and production deployment via LangServe. Use when working with LangChain or comparing LLM application frameworks.
magnus919/agent-skills/langgraph Skill
169★ repoBuild multi-agent AI systems with LangGraph — the low-level orchestration framework for stateful, graph-based agent workflows. Covers supervisor, swarm, and hierarchical multi-agent patterns; subgraph composition; state management (checkpointers/stores); persistence; evals; and production debugging. Reach for this when designing agent architectures that need cycles, conditional branching, parallel execution, or human-in-the-loop patterns.
magnus919/agent-skills/llamaindex Skill
169★ repoExpert skill for building LLM applications with the LlamaIndex framework — RAG pipelines, multi-agent orchestration, event-driven workflows, knowledge graph construction, production deployment, and evaluation. Use when working with LlamaIndex or comparing RAG and agent orchestration frameworks.
magnus919/agent-skills/open-knowledge-format Skill
169★ repoGoogle's Open Knowledge Format (OKF) v0.1 — an open, vendor-neutral spec for representing knowledge as markdown files with YAML frontmatter, designed for AI agent consumption. Use when the user mentions OKF, Open Knowledge Format, Google's knowledge format, LLM wiki bundles, agent knowledge packs, creating OKF bundles, validating OKF documents, or converting knowledge into the OKF standard.
magnus919/agent-skills/pydanticai Skill
169★ repoBuild type-safe AI agents and graph-based workflows with PydanticAI and PydanticGraph. Agent creation, function tools, capabilities, dependency injection, structured output, streaming, multi-agent patterns, testing, evals, and graph state machines. Use whenever you are building agents, tool-using LLM workflows, or graph-based state machines in Python.
magnus919/agent-skills/spec-driven-development Skill
169★ repoSpec-Driven Development (SDD) methodology for AI software factories — where structured specifications are the input, AI agents generate the code, and quality gates enforce correctness at each pipeline phase. Use when designing a spec → review → decompose → implement → verify pipeline that any AI coding tool (Claude Code, Cursor, Hermes Agent, Devin, OpenHands) can follow.
Junliu1066/vibe-coding-kit/skills/vibe-coding-prd Skill
167★ repo一个"访谈式"的需求 Agent:通过一问一答 + 在岔路口给建议,把一个说不清需求的人,带到一份 AI 能直接照着开工的 PRD。 当用户说"我想做个 XX 但不知道怎么跟 AI 说"、"帮我理需求"、"帮我写个 PRD"、"我有个想法,帮我想清楚再开工", 或在 Claude Code / Codex 里准备开一个新项目、却还没把要做什么定清楚时,使用此 Skill。 它不是让用户填表格,而是像产品顾问一样访谈:一次只问一件事,遇到用户判断不了的选择(尤其"产品形态"和"技术栈") 会给出 2-3 个大白话选项 + 推荐一个最简可行的默认 + 一句话讲清代价,等用户拍板。 访谈结束后,直接把 prd.md 和 项目说明书.md 写进项目目录,并提议生成可点击的 HTML 原型。 这是 vibe-coding-kit 套件里"把需求从人嘴里问出来"的交互入口。
Impertio-Studio/Frappe_Claude_Skill_Package/skills/source/agents/frappe-agent-interpreter Skill
no license167★ repoUse when receiving vague or unclear ERPNext/Frappe development requests that need interpretation. Transforms requirements like 'make invoice auto-calculate' or 'add approval workflow' into concrete technical specifications. Determines which Frappe mechanisms to use and maps to the full 61-skill catalog. Keywords: vague requirement, clarify scope, translate business need, technical spec, implementation plan, what does this mean, unclear requirement, translate to code, how to build this.
Impertio-Studio/Frappe_Claude_Skill_Package/skills/source/agents/frappe-agent-migrator Skill
no license167★ repoUse when migrating a Frappe app between major versions, detecting breaking API changes, or resolving post-migration errors. Prevents failed migrations from undetected deprecated APIs, removed methods, and changed function signatures. Covers breaking change detection v14-v15-v16, deprecated API mapping, migration checklist, common migration errors, automatic fix suggestions. Keywords: migration, version upgrade, breaking changes, deprecated API, v14, v15, v16, migrate, compatibility, upgrade ERPNext, version change breaks, after update errors, deprecated method..
Impertio-Studio/Frappe_Claude_Skill_Package/skills/source/agents/frappe-agent-validator Skill
no license167★ repoUse when reviewing or validating Frappe/ERPNext code against best practices and common pitfalls. Checks generated code before deployment, validates against all 61 frappe-* skills, catches v16 patterns (extend_doctype_class, type annotations), validates ops patterns (bench commands, deployment), and generates correction reports. Keywords: review code, check script, validate deployment, find bugs, code quality, check my code, is this correct, code review, before deploying, best practices check.
aahl/skills/skills/mcp-deepwiki Skill
166★ repoSkills for accessing and searching docs in DeepWiki/GitHub’s public code repositories can help users understand open-source project source codes, and users can also ask questions directly about the code docs.
The same file is in 6 other repositories
aahl/skills/skills/mcp-duckgo Skill
166★ repoSkills for web search and content scraping via DuckDuckGo MCP Server. Used when users need online searching and web scraping.
The same file is in 3 other repositories
aahl/skills/skills/mcp-hass Skill
166★ repoThe skill for control Home Assistant smart home devices and query states using MCP protocol.
The same file is in 3 other repositories
aahl/skills/skills/mcp-lark Skill
166★ repoBased on FeiShu(飞书) / Lark's OpenAPI MCP server, manage user information, chats, emails, cloud documents, multidimensional tables, tasks, calendars, etc.
The same file is in 3 other repositories
aahl/skills/skills/mcp-vods Skill
166★ repo用于追剧/追番的技能,为AI提供搜索影视播放地址的能力,并支持在小米电视上直接播放。当用户想搜索影视、动漫、短剧、综艺等节目信息或更新进度时使用此技能。
The same file is in 3 other repositories
TabooHarmony/roblox-brain/skills/roblox-studio-mcp Skill
166★ repoUse when working with Roblox Studio through built-in MCP for scripts, scenes, generated assets, input, or playtesting.
The same file is in 1 other repository
kangarooking/system-prompt-skills/context-management Skill
162★ repo当系统提示词需要设计 token 预算分配、上下文压缩策略、延迟加载机制、记忆持久化方案时调用此 Skill。适用于长对话 AI 助手、代码编辑器集成、研究型 Agent、多会话系统等需要精细管理上下文窗口的场景。不适用于:单轮交互系统(无上下文管理需求)、纯无状态 API(无对话历史)、简单的 prompt 模板设计。当需求聚焦于"如何搜索外部信息"而非"如何管理已有信息"时,应该用 search-integration 而非本 Skill。
kangarooking/system-prompt-skills/memory-system Skill
162★ repo当需要为 AI 设计记忆存储、检索、应用和更新机制时调用此 skill。典型场景包括:设计持久化记忆架构(用户偏好、历史上下文、项目知识)、定义记忆的创建/读取/更新/删除生命周期、实现静默记忆应用(不在回复中透露记忆内容)、管理敏感记忆边界。 不适用于:定义工具接口(tool-specification)、定义安全规则(safety-guardrails)、定义人格风格(personality-system)。 关键 trigger 信号:AI 需要跨会话记住用户信息、记忆内容可能敏感、需要在回复中隐式应用记忆而非显式引用、用户要求"记住这个"。
kangarooking/system-prompt-skills/persona-design Skill
162★ repo当需要为 AI 产品定义核心身份、角色声明和能力边界时调用此 skill。典型场景包括:设计新 AI 产品的 system prompt 首段、为不同场景创建差异化角色(如教学助手 vs 编程代理)、重新定义 AI 与用户的关系框架。 不适用于:纯人格风格调优(应使用 personality-system)、安全规则制定(应使用 safety-guardrails)、工具集成(应使用 tool-specification)。 关键 trigger 信号:产品需要 AI 有明确自我认知、多个角色共享底层能力但身份不同、需要限制 AI 在特定领域内运作、用户会问"你是谁"。
alexgreensh/repo-forensics/skills/forensify Skill
no license161★ repoCross-agent self-inspection of your AI-agent stack. Audits skills, MCP servers, hooks, plugins, commands, credentials, and memory files across Claude Code, Codex, OpenClaw, and NanoClaw. Produces a structured inventory and narrative briefing with cross-ecosystem risk analysis. Use when the user asks to audit their own setup, check what they have installed, review their agent stack security posture, or understand cross-tool interactions. Use when a user has accumulated skills/plugins/MCP servers over time and wants visibility into their attack surface. Use after installing new skills or plugins. Do NOT use for vetting external code before install (that is repo-forensics). Do NOT use for incident response during active attacks. Do NOT use for fixing or patching vulnerabilities (forensify is read-only).
zach22-1999/amazon-skills/skills/zach-seller-skill-creator Skill
159★ repo亚马逊卖家专用的 skill 创建器(中文)。当用户想把一个亚马逊运营/自媒体/日常工作流程变成可复用的 skill 时使用。触发场景包括但不限于:用户说"我想做一个 skill""把这个流程变成 skill""帮我写个自动化""优化我已有的 skill""给这个工作流做个自动化",即使用户没用"skill"这个词,只要在描述"以后每次都这样做"的重复性工作时也应触发。本 skill 的核心差异:强制用户先回答 6 个业务问题(业务目标/过去做法/具体步骤/方法论/调用方式/期望输出)再进入创建流程,防止产出空洞 skill。Create new skills, improve existing skills, run evals and benchmarks — tailored for Amazon sellers with a Chinese-first workflow.
The same file is in 2 other repositories
EXboys/evotown/backend/arena_skills/skill-creator Skill
no license157★ repoCreate or update AgentSkills. Use when designing, structuring, or packaging skills with scripts, references, and assets.
EXboys/evotown/integrations/openclaw/evotown/skills/evotown-handoff Skill
no license157★ repoQueue a cross-agent or cross-team handoff through the Evotown control plane.
apache/magpie/skills/optimize-skill Skill
145★ repoOptimize an existing framework skill (or sweep a set of them) by applying the restructuring patterns proven on the security-skill suite: split an oversized `SKILL.md` into linked sibling docs, lift concrete/project-specific values out of the body into `<project-config>` placeholders, replace in-agent-context body reads with out-of-context tool calls, batch per-item fetches into a single upfront pass, and add a deterministic pre-flight no-op classifier ahead of LLM passes. Every change is a behavior- preserving proposal the maintainer signs off on; the skill validator must stay green before and after. The refactoring sibling of `write-skill` (which authors net-new skills).
apache/magpie/skills/pairing-multi-agent-review Skill
145★ repoFan a local diff through three independent, axis-focused review passes (correctness, security, conventions), then merge the findings into a single structured report. Each pass is isolated so findings from one axis cannot suppress or bias the others. The merged report uses the same format as pairing-self-review so the developer gets a consistent signal regardless of which Agentic Pairing skill they invoke.
apache/magpie/skills/setup-isolated-setup-install Skill
145★ repoGuide an adopter through the first-time install of the framework's secure agent setup (bubblewrap + socat + claude-code, sandbox/permissions/clean-env layers). Walks every step interactively; never auto-runs sudo, shell-rc edits, or settings overwrites.
arbiterForge/codeArbiter/core/surface/skills/dispatching-parallel-agents Skill
141★ repoWhen you can't trust yourself with your code base, trust Arbiter.
arbiterForge/codeArbiter/core/surface/skills/executing-plans Skill
141★ repoWhen you can't trust yourself with your code base, trust Arbiter.
arbiterForge/codeArbiter/core/surface/skills/secret-handling Skill
141★ repoWhen you can't trust yourself with your code base, trust Arbiter.
arbiterForge/codeArbiter/core/surface/skills/subagent-driven-development Skill
141★ repoWhen you can't trust yourself with your code base, trust Arbiter.
arbiterForge/codeArbiter/core/surface/skills/using-git-worktrees Skill
141★ repoWhen you can't trust yourself with your code base, trust Arbiter.
arbiterForge/codeArbiter/plugins/ca-codex/routines/dispatching-parallel-agents Skill
141★ repoThe parallel fan-out primitive. Routed to by any skill or command that splits work across independent units and dispatches an agent per unit — subagent-driven-development, /sprint, parallel /review. It owns the dispatch/collect/funnel discipline: bound concurrency, isolate units, collect every result, dedupe overlap, and funnel through finding-triage then checkpoint-aggregator. Raw agent output is never consumed before the funnel runs; an agent that errors drops its unit without corrupting the batch.
arbiterForge/codeArbiter/plugins/ca-codex/routines/executing-plans Skill
141★ repoThe checkpoint coordinator for /feature. Routed to by /feature once a writing-plans plan exists. Groups tasks into batches, delegates each batch to subagent-driven-development (fresh author agent per task, full review chain, fresh verification), then stops for a human checkpoint before the next batch. The checkpointed counterpart to /sprint's autonomous run.
arbiterForge/codeArbiter/plugins/ca-codex/routines/secret-handling Skill
141★ repoThe secret-source gate. Routed to when changed code reads, writes, or passes a secret — API key, token, password, connection string, signing key, certificate, or any value that grants access. Validates that every secret comes from the approved store and never lands in source, log, test fixture, error response, image, or LLM prompt. The auth-crypto-reviewer agent is dispatched as the reviewer.
arbiterForge/codeArbiter/plugins/ca-codex/routines/subagent-driven-development Skill
141★ repoThe implementation engine. Routed to by /sprint (full plan, autonomous) and by executing-plans (scoped batch, checkpoint-gated). One fresh subagent per task — test-first via tdd — followed by spec-compliance review, quality review, and fresh-run verification. No single context accumulates drift, and nothing is accepted on a subagent's word.
arbiterForge/codeArbiter/plugins/ca-codex/routines/using-git-worktrees Skill
141★ repoOPTIONAL per-task isolation for autonomous parallel work. Routed to only on explicit opt-in by subagent-driven-development or dispatching-parallel-agents, so parallel units mutate files without colliding. Stands up one worktree per unit, works it in isolation, then integrates each unit back onto the caller's working branch for the caller's single commit-gate + finishing-a-development-branch exit. Never on the default path; it does not bypass a gate or finish per unit.
arbiterForge/codeArbiter/plugins/ca-pi/routines/dispatching-parallel-agents Skill
141★ repoThe parallel fan-out primitive. Routed to by any skill or command that splits work across independent units and dispatches an agent per unit — subagent-driven-development, /sprint, parallel /review. It owns the dispatch/collect/funnel discipline: bound concurrency, isolate units, collect every result, dedupe overlap, and funnel through finding-triage then checkpoint-aggregator. Raw agent output is never consumed before the funnel runs; an agent that errors drops its unit without corrupting the batch.
arbiterForge/codeArbiter/plugins/ca-pi/routines/executing-plans Skill
141★ repoThe checkpoint coordinator for /feature. Routed to by /feature once a writing-plans plan exists. Groups tasks into batches, delegates each batch to subagent-driven-development (fresh author agent per task, full review chain, fresh verification), then stops for a human checkpoint before the next batch. The checkpointed counterpart to /sprint's autonomous run.
arbiterForge/codeArbiter/plugins/ca-pi/routines/secret-handling Skill
141★ repoThe secret-source gate. Routed to when changed code reads, writes, or passes a secret — API key, token, password, connection string, signing key, certificate, or any value that grants access. Validates that every secret comes from the approved store and never lands in source, log, test fixture, error response, image, or LLM prompt. The auth-crypto-reviewer agent is dispatched as the reviewer.
arbiterForge/codeArbiter/plugins/ca-pi/routines/subagent-driven-development Skill
141★ repoThe implementation engine. Routed to by /sprint (full plan, autonomous) and by executing-plans (scoped batch, checkpoint-gated). One fresh subagent per task — test-first via tdd — followed by spec-compliance review, quality review, and fresh-run verification. No single context accumulates drift, and nothing is accepted on a subagent's word.
arbiterForge/codeArbiter/plugins/ca-pi/routines/using-git-worktrees Skill
141★ repoOPTIONAL per-task isolation for autonomous parallel work. Routed to only on explicit opt-in by subagent-driven-development or dispatching-parallel-agents, so parallel units mutate files without colliding. Stands up one worktree per unit, works it in isolation, then integrates each unit back onto the caller's working branch for the caller's single commit-gate + finishing-a-development-branch exit. Never on the default path; it does not bypass a gate or finish per unit.
arbiterForge/codeArbiter/plugins/ca-sandbox/skills/sandbox-claude-inside Skill
141★ repoRun Claude Code INSIDE a ca-sandbox box (`--with-claude`). Routed to when the user wants an agent loop running against an isolated, ephemeral sandbox rather than the host. Authenticates via an env-injected CLAUDE_CODE_OAUTH_TOKEN with no host bind of ~/.claude; the image pins the CLI and disables the autoupdater; HOME is backed by a named volume so the .claude state persists across restart. Five gated phases — posture, image, token, run, teardown. The hard default is offline or Anthropic-domains-only egress, and the token volume is NEVER co-mounted with an untrusted-code run; both are enforced, not advised.
arbiterForge/codeArbiter/plugins/ca/skills/dispatching-parallel-agents Skill
141★ repoThe parallel fan-out primitive. Routed to by any skill or command that splits work across independent units and dispatches an agent per unit — subagent-driven-development, /sprint, parallel /review. It owns the dispatch/collect/funnel discipline: bound concurrency, isolate units, collect every result, dedupe overlap, and funnel through finding-triage then checkpoint-aggregator. Raw agent output is never consumed before the funnel runs; an agent that errors drops its unit without corrupting the batch.
arbiterForge/codeArbiter/plugins/ca/skills/executing-plans Skill
141★ repoThe checkpoint coordinator for /feature. Routed to by /feature once a writing-plans plan exists. Groups tasks into batches, delegates each batch to subagent-driven-development (fresh author agent per task, full review chain, fresh verification), then stops for a human checkpoint before the next batch. The checkpointed counterpart to /sprint's autonomous run.
arbiterForge/codeArbiter/plugins/ca/skills/secret-handling Skill
141★ repoThe secret-source gate. Routed to when changed code reads, writes, or passes a secret — API key, token, password, connection string, signing key, certificate, or any value that grants access. Validates that every secret comes from the approved store and never lands in source, log, test fixture, error response, image, or LLM prompt. The auth-crypto-reviewer agent is dispatched as the reviewer.
arbiterForge/codeArbiter/plugins/ca/skills/subagent-driven-development Skill
141★ repoThe implementation engine. Routed to by /sprint (full plan, autonomous) and by executing-plans (scoped batch, checkpoint-gated). One fresh subagent per task — test-first via tdd — followed by spec-compliance review, quality review, and fresh-run verification. No single context accumulates drift, and nothing is accepted on a subagent's word.
arbiterForge/codeArbiter/plugins/ca/skills/using-git-worktrees Skill
141★ repoOPTIONAL per-task isolation for autonomous parallel work. Routed to only on explicit opt-in by subagent-driven-development or dispatching-parallel-agents, so parallel units mutate files without colliding. Stands up one worktree per unit, works it in isolation, then integrates each unit back onto the caller's working branch for the caller's single commit-gate + finishing-a-development-branch exit. Never on the default path; it does not bypass a gate or finish per unit.
Dave-London/Pare/io.github.Dave-London/http MCP server
141★ repoMCP server for HTTP requests (curl) with structured, token-efficient output
Dave-London/Pare/io.github.Dave-London/make MCP server
141★ repoMCP server for Make/Just task runners with structured, token-efficient output
Dave-London/Pare/io.github.Dave-London/github MCP server
141★ repoMCP server for GitHub operations (PRs, issues, actions) with structured, token-efficient output
Dave-London/Pare/io.github.Dave-London/search MCP server
141★ repoMCP server for code search (ripgrep + fd) with structured, token-efficient output
jianshuo/claude-skills/wjs-evaling-voicedrop-prompts Skill
140★ repoUse when 王建硕 wants to evaluate whether a change to VoiceDrop's 挖矿 system prompt is actually better than the live version — runs the local eval harness (golden fixtures × champion-vs-candidate, same input), dispatches blind pairwise judge subagents, aggregates a win-rate verdict, and on approval promotes the candidate into agent/src/prompts/mine.js. Triggers — "评估 prompt"、"挖矿 prompt 改好了吗"、"eval prompt"、"比一比两版 prompt"、"/wjs-evaling-voicedrop-prompts".
robertguss/claude-code-toolkit/plugins/compound-writing/skills/writing-orchestration Skill
139★ repoThis skill should be used when orchestrating complex writing workflows with multiple phases. It provides two-agent orchestration patterns, the two-gate content readiness assessment, 10 baseline writing strategies, 20+ situational strategies, and quality checkpoints. Inspired by the Spiral Writing System.
The same file is in 1 other repository
robertguss/claude-code-toolkit/skills/handoff Skill
139★ repoA collection of custom skills that extend Claude's capabilities with specialized workflows, methods, and domain knowledge.
The same file is in 1 other repository
karaage0703/ai-assistant-workspace/skills/xs-multi-agent Skill
136★ repo異なるproviderのAI(Codex/Claude/Grok/Cursor/Antigravity等)から独立意見を集め、共通論点で相互検討して統合する。最初に利用可能エージェントを検出する。「みんなで考えて」「マルチエージェントで分析して」「複数のAIでレビューして」と言われたら使う。
karaage0703/ai-assistant-workspace/skills/xs-skill-creator Skill
136★ repoスキルの作成・改善を行うスキル。既存スキルの分析から抽出したパターンとテンプレートで統一感のあるスキルを効率的に作成。スキルを作りたい、新しいスキルを追加したい、スキルを改善したい、SKILL.mdを書きたい場合に使用。