Context & AI engineering
26,304 rows from 10,117 repositories
Making an agent good at your problem. The thing this whole site is about.
- Prompts1,602
- Agents & orchestration3,106
- Memory & context2,410
- Evals & benchmarks944
- RAG & retrieval912
- MCP & tools8,994
- Writing skills1,028
What Context & AI engineering skills agree on
1,865 skill files read, by 1,328 of the 2,349 authors on this shelf whose files we hold, 2026-09-06
The middle one of the 226 measured here is ~1.7k tokens long, counted with cl100k_base
Counted by distinct author, so one author publishing three of these counts once. Where a claim sits in fewer files than authors, that is said: a claim held by forty authors across three files is one file people copied, not forty people who agreed. Near-identical wordings are grouped and the other wordings are shown, so the grouping is yours to check.
What they tell the agent to do
- Dispatch a fresh subagent for each task76 of 1328 in 59 filesalso worded as dispatch a fresh implementer subagent for each task; dispatch a fresh implementer subagent per task
- Perform spec compliance review before code quality review44 of 1328 in 34 filesalso worded as perform spec compliance review first; Perform spec compliance review after each task
- Dispatch a final code reviewer after all tasks38 of 1328 in 26 filesalso worded as dispatch final reviewer after all tasks finish; use the final code reviewer after all tasks finish
- Answer subagent questions before allowing implementation36 of 1328 in 26 filesalso worded as Answer all subagent questions before implementation; answer all subagent questions before they begin work
- Use the least powerful model capable of the task33 of 1328 in 26 filesalso worded as use the least powerful model for mechanical tasks; use the least powerful model suitable for the task
- Create a TodoWrite list for all tasks32 of 1328 in 22 filesalso worded as create a todo list for all tasks; Create a TodoWrite list
- Perform a task review after each implementation31 of 1328 in 24 filesalso worded as perform a task review after each task; Perform code quality review after each task
- Extract all tasks and context from the plan29 of 1328 in 20 filesalso worded as extract all tasks and context from plan; extract all tasks and context before starting
- Provide full task text to subagents28 of 1328 in 20 filesalso worded as provide full task text and context to subagents; provide full task text and context to each subagent
- Use git worktrees for isolated workspaces25 of 1328 in 20 filesalso worded as use git worktrees for isolation; use git worktrees to ensure isolated workspaces
- Specify the model explicitly when dispatching a subagent23 of 1328 in 18 filesalso worded as explicitly specify the model when assigning subagents; specify models explicitly for each subagent
- Execute all tasks from the plan without stopping21 of 1328 in 16 filesalso worded as execute all tasks in the plan without pausing; execute all tasks in the plan without pausing for approval
- Verify robots.txt allows AI crawlers21 of 1328 in 7 filesalso worded as Verify AI bot access in robots.txt; Verify AI crawler accessibility and robots policies
- Monitor AI visibility monthly20 of 1328 in 6 filesalso worded as Monitor AI visibility and citation rates monthly
- Read product marketing context before starting20 of 1328 in 6 filesalso worded as Read product marketing context before asking questions; check product marketing context before starting
What they tell it not to do
- Do not dispatch multiple implementation subagents in parallel70 of 1328 in 53 filesalso worded as Dispatch multiple implementation subagents in parallel; never dispatch multiple implementation subagents in parallel
- Do not skip task reviews58 of 1328 in 42 filesalso worded as skip review loops; never skip review loops
- Do not make subagents read the plan file53 of 1328 in 41 filesalso worded as Do not make subagents read the plan file; never make subagents read the plan file
- Do not ignore subagent questions44 of 1328 in 32 filesalso worded as never ignore subagent questions; ignore subagent questions
- Never proceed with unfixed issues41 of 1328 in 27 filesalso worded as Do not proceed with unfixed issues; proceed with unfixed issues
- Do not start code quality review before spec compliance40 of 1328 in 30 filesalso worded as never start code quality review before spec compliance; never skip spec compliance or code quality reviews
- Do not start implementation on main branch without consent38 of 1328 in 32 filesalso worded as never start implementation on main branch without consent; do not use the main or master branch for implementation
- Never accept close enough on spec compliance34 of 1328 in 24 filesalso worded as do not accept close enough on spec compliance; accept close enough on spec compliance
- Never let self-review replace actual review31 of 1328 in 21 filesalso worded as let self-review replace external review; do not let implementer self-review replace actual review
- Do not pause to check in with human partner between tasks29 of 1328 in 24 filesalso worded as do not pause to check in with human partner; do not pause for human check-ins between tasks
What they expect to be installed
- git253 of 1328
- bash129 of 1328
- grep122 of 1328
- curl91 of 1328
- python390 of 1328
- claude83 of 1328 in 79 files
- node82 of 1328
- gh74 of 1328
- jq67 of 1328
- npm66 of 1328
What they ask it to produce
- SKILL.md47 of 1328also worded as SKILL.md file; draft SKILL.md
- Implemented code35 of 1328 in 25 filesalso worded as Adapted implementation code; implemented agent code
- CLAUDE.md33 of 1328also worded as Updated CLAUDE.md; CLAUDE.md updates
- Pull requests31 of 1328 in 24 filesalso worded as pull request; one pull request
- TodoWrite30 of 1328 in 20 filesalso worded as TodoWrite list
- AGENTS.md27 of 1328also worded as AGENT.md; AGENTS.md.bak
- Markdown report26 of 1328 in 24 filesalso worded as Markdown security report; markdown audit report
- AI visibility audit report24 of 1328 in 7 files
When Context & AI engineering authors say to reach for one
The situations these authors wrote into their own files, counted out of the same 1,328 authors, with the skills that name each one
- Plan tasks are mostly or entirely independent65 of 1328 in 48 files
- Subagent driven development
- Sp subagent driven development
- A7b8c9d0 e1f2 4a5b 9c8d 7e8f9a0b1c2d
- Subagent driven development skills cycleaddict subagent driven development
- Subagent driven development skills cygnusfear subagent driven development
- Subagent driven development skills davila7 subagent driven development
- and 52 more on this shelf
- Have or exist a written implementation plan62 of 1328 in 47 files
- Subagent driven development
- A7b8c9d0 e1f2 4a5b 9c8d 7e8f9a0b1c2d
- Subagent driven development skills cycleaddict subagent driven development
- Subagent driven development skills cygnusfear subagent driven development
- Subagent driven development skills davila7 subagent driven development
- Subagent driven development skills obra subagent driven development
- and 51 more on this shelf
- Staying in the current session56 of 1328 in 41 files
- Subagent driven development
- Sp subagent driven development
- A7b8c9d0 e1f2 4a5b 9c8d 7e8f9a0b1c2d
- Subagent driven development skills cycleaddict subagent driven development
- Subagent driven development skills cygnusfear subagent driven development
- Subagent driven development skills davila7 subagent driven development
- and 52 more on this shelf
- User asks to create an agent50 of 1328 in 48 files
- User asks to use Claude API34 of 1328 in 26 files
- Claude coach
- Claude api
- and 7 more on this shelf
- When starting a session18 of 1328 in 13 files
- Using agent skills
- Claude mem
- and 1 more on this shelf
- Setting up a new project15 of 1328
How Context & AI engineering skills are built
17,957 skill directories by 3,594 authors, read from their repositories’ own file trees 2026-08-05
The middle bundle among those shipping files is 3 files, 19.4 KB beside SKILL.md
Counted by distinct author, same as above, so one author publishing forty template copies counts once. SKILL.md itself is not counted as a file, so a single-file skill is one where that file is the whole skill.
The shape
- SKILL.md is the whole skill1,002 of 3,594 authors, 8,446 of 17,957 skills
- files ship beside it2,592 of 3,594 authors, 9,511 of 17,957 skills
- executable scripts ship inside1,381 of 3,594 authors, 3,721 of 17,957 skills
The folders they converge on
- references/1,629 of 3,594 authors, 4,820 of 17,957 skills
- scripts/1,138 of 3,594 authors, 3,159 of 17,957 skills
- agents/689 of 3,594 authors, 1,334 of 17,957 skills
- assets/491 of 3,594 authors, 941 of 17,957 skills
- templates/261 of 3,594 authors, 431 of 17,957 skills
- examples/238 of 3,594 authors, 413 of 17,957 skills
getsentry/sentry-mcp/io.github.getsentry/sentry-mcp MCP server
in 2 stacksno license808★ repoMCP server for Sentry - error monitoring, issue tracking, and debugging for AI assistants
anthropics/skills/skills/mcp-builder Skill
in 1 stackno license168,934★ repoPublic repository for Agent Skills
microsoft/playwright-mcp/io.github.microsoft/playwright-mcp MCP server
in 1 stack36,126★ repoPlaywright Tools for MCP
Serena MCP: the IDE for your agent
oraios/serena/io.github.oraios/serena MCP server
in 1 stack27,951★ repoA powerful toolkit for coding, providing semantic retrieval and editing capabilities.
GLips/Figma-Context-MCP/io.github.GLips/Figma-Context-MCP MCP server
in 1 stack15,643★ repoGive your coding agent access to your Figma data. Implement designs in any framework in one-shot.
anthropics/skills/skills/skill-creator Skill
in 1 stackno license168,934★ repoPublic repository for Agent Skills
D4Vinci/Scrapling/io.github.D4Vinci/Scrapling MCP server
73,239★ repoWeb scraping with stealth HTTP, real browsers, and Cloudflare bypass. CSS selectors supported.
ruvnet/claude-flow/io.github.ruvnet/claude-flow MCP server
67,334★ repoAI orchestration with hive-mind swarms, neural networks, and 87 MCP tools for enterprise dev.
shanraisshan/claude-code-best-practice/.claude/skills/agent-browser Skill
64,157★ repofrom vibe coding to agentic engineering - practice makes claude perfect
ChromeDevTools/chrome-devtools-mcp/io.github.ChromeDevTools/chrome-devtools-mcp MCP server
48,992★ repoMCP server for Chrome DevTools
HeyPuter/puter/com.puter/mcp-server MCP server
43,030★ repoPuter MCP enables AI tools to interact with Puter: manage files, websites, workers, and more
DeusData/codebase-memory-mcp/io.github.DeusData/codebase-memory-mcp MCP server
39,060★ repoCodebase knowledge graph for AI agents — 159 languages, sub-ms queries, 99% fewer tokens.
bytedance/UI-TARS-desktop/io.github.bytedance/mcp-server-filesystem MCP server
38,554★ repoMCP server for filesystem access
github/awesome-copilot/.github/skills/agentic-workflows Skill
37,897★ repoCommunity-contributed instructions, agents, skills, and configurations to help you make the most of GitHub Copilot.
PostHog/posthog/io.github.PostHog/mcp MCP server
no license37,662★ repoOfficial PostHog MCP Server for product analytics, feature flags, experiments, and more.
nanocoai/nanoclaw/.claude/skills/add-atomic-chat-tool Skill
30,508★ repoA lightweight alternative to OpenClaw that runs in containers for security. Connects to WhatsApp, Telegram, Slack, Discord, Gmail and other messaging apps,, has memory, scheduled jobs, and runs directly on Anthropic's Agents SDK
agentskills/agentskills/ai.com.mcp/skills-search MCP server
24,236★ repoSearch and discover Agent Skills from the skills.sh registry. Powered by HAPI MCP server.
alibaba/open-code-review/plugins/open-code-review/skills/open-code-review-delegate Skill
20,413★ repoOpen-source & free — Battle-tested at Alibaba's scale. Hybrid architecture code review tool: deterministic pipelines + LLM Agent, precise line-level comments, built-in fine-tuned ruleset (NPE, thread-safety, XSS, SQL injection), OpenAI & Anthropic compatible.
KKKKhazix/khazix-skills/aihot Skill
19,372★ repo数字生命卡兹克开源的 AI Skills 合集 | Agent Skills: leader(帮你定义目标), neat-freak 洁癖, hv-analysis, khazix-writer & more — Claude Code, Codex & 40+ agents
googleapis/mcp-toolbox/io.github.googleapis/mcp-toolbox MCP server
16,180★ repoMCP Toolbox for Databases enables your agent to connect to your database.
googleapis/genai-toolbox/io.github.googleapis/genai-toolbox MCP server
16,178★ repoMCP Toolbox for Databases enables your agent to connect to your database.
microsoft/SkillOpt/plugins/claude-code/skills/skillopt-sleep Skill
15,937★ repoSkillOpt is a text-space optimizer that trains reusable natural-language skills for frozen LLM agents through trajectory-driven edits, validation-gated updates, and deployable best_skill.md artifacts.
kubeshark/kubeshark/io.github.kubeshark/mcp MCP server
12,042★ repoReal-time Kubernetes network traffic visibility and API analysis for HTTP, gRPC, Redis, Kafka, DNS.
Orchestra-Research/AI-Research-SKILLs/0-autoresearch-skill Skill
11,505★ repoComprehensive open-source library of AI research and engineering skills for any AI model. Package the skills and your claude code/codex/gemini agent will be an AI research agent with full horsepower. Maintained by Orchestra Research.
nidhinjs/prompt-master Skill
11,108★A Claude skill that writes the accurate prompts for any AI tool. Zero tokens or credits wasted. Full context and memory retention
wonderwhy-er/DesktopCommanderMCP/io.github.wonderwhy-er/desktop-commander MCP server
9,328★ repoMCP server for terminal commands, file operations, and process management
MystenLabs/sui/io.github.jessiemongeon1/sui-kapa MCP server
7,735★ repoSemantic search across Sui documentation and knowledge sources, powered by kapa.ai
firecrawl/firecrawl-mcp-server/io.github.firecrawl/firecrawl-mcp-server MCP server
7,233★ repoMCP server for Firecrawl — search, scrape, and interact with the web.
modelcontextprotocol/registry/ai.com.mcp/registry MCP server
no license7,125★ repoPublish and discover MCP servers via the official MCP Registry. Powered by HAPI MCP server.
CursorTouch/Windows-MCP/io.github.CursorTouch/Windows-MCP MCP server
6,733★ repoAn MCP Server for computer-use in Windows OS
trailofbits/skills/plugins/agentic-actions-auditor/skills/agentic-actions-auditor Skill
6,559★ repoTrail of Bits Claude Code skills for security research, vulnerability detection, and audit workflows
airweave-ai/airweave/io.github.airweave-ai/search MCP server
6,542★ repoMCP server for searching Airweave collections with natural language queries.
nirholas/claude-code/io.github.nirholas/claude-code-explorer-mcp MCP server
no license6,252★ repoExplore the Claude Code CLI source — browse tools, commands, search code, and more.
mobile-next/mobile-mcp/io.github.mobile-next/mobile-mcp MCP server
5,927★ repoMCP server for iOS and Android Mobile Development, Automation and Testing
Klavis-AI/klavis/ai.klavis/strata MCP server
5,791★ repoMCP server for progressive tool usage at any scale (see https://klavis.ai)
WenyuChiou/awesome-agentic-ai-zh/examples/stage-5/tool-calling-tutor Skill
5,144★ repoA trilingual (繁中 / English / 简中) learning roadmap for agentic AI: from LLM basics to multi-agent systems, with 240+ curated resources and hands-on examples. 中文 AI agent 學習地圖。
dotnet/skills/.agents/skills/authoring-github-workflows Skill
5,132★ repoRepository for skills to assist AI coding agents with .NET and C#
KnockOutEZ/wigolo/io.github.KnockOutEZ/wigolo MCP server
no license4,534★ repoLocal-first web intelligence MCP server for AI coding agents
firebase/firebase-tools/io.github.firebase/firebase-mcp MCP server
4,451★ repoGives AI development tools Firebase-specific capabilities and expertise.
homeassistant-ai/ha-mcp/io.github.homeassistant-ai/ha-mcp MCP server
4,336★ repoComprehensive Model Context Protocol server for managing Home Assistant through AI assistants.
antvis/mcp-server-chart/io.github.antvis/mcp-server-chart MCP server
4,312★ repoA Model Context Protocol server for generating charts using AntV.
wuyoscar/GPT-Image2-Skill/skills/gpt-image Skill
4,246★ repoGPT Image 2 prompt gallery, image prompt library, agentic skill, and CLI for OpenAI image generation/editing
cloudflare/mcp-server-cloudflare/com.cloudflare.mcp/mcp MCP server
4,073★ repoCloudflare MCP servers
eugeniughelbur/obsidian-second-brain Skill
4,022★Persistent memory for Claude Code and 6 other CLI agents, stored as plain markdown in your Obsidian vault. Stop re-explaining your projects, decisions and people every session. 45 commands: hybrid semantic search, self-rewriting notes, key-less web research, and scheduled agents that maintain the vault while you sleep.
callstackincubator/agent-device/io.github.callstackincubator/agent-device MCP server
4,021★ repoLet AI agents inspect, control, and debug real iOS, Android, desktop, and TV apps
parcadei/Continuous-Claude-v3/.claude/plugins/braintrust-tracing/skills/trace-claude-code Skill
3,882★ repoContext management for Claude Code. Hooks maintain state via ledgers and handoffs. MCP execution without context pollution. Agent orchestration with isolated context windows.
IvanMurzak/Unity-MCP/io.github.IvanMurzak/Unity-MCP MCP server
3,845★ repoMake 3D games in Unity Engine with AI. MCP Server + Plugin for Unity Editor and Unity games.
gotalab/cc-sdd/.agents/skills/cc-sdd-new-agent Skill
3,620★ repoTurn approved specs into long-running autonomous implementation. A minimal, adaptable SDD harness with Agent Skills for Claude Code, Codex, Cursor, Copilot, Windsurf, OpenCode, Gemini CLI, and Antigravity.
basicmachines-co/basic-memory/io.github.basicmachines-co/basic-memory MCP server
3,612★ repoLocal-first knowledge management with bi-directional LLM sync via Markdown files.
microsoft/mcp/com.microsoft/azure MCP server
3,553★ repoAll Azure MCP tools to create a seamless connection between AI agents and Azure services.
revolist/revogrid/io.github.revolist/revogrid-mcp MCP server
3,434★ repoHosted MCP server for RevoGrid DataGrid docs, examples, feature checks, and migration guidance.
browserbase/mcp-server-browserbase/ai.smithery/browserbasehq-mcp-browserbase MCP server
3,416★ repoProvides cloud browser automation capabilities using Stagehand and Browserbase, enabling LLMs to i…
bytebase/dbhub/io.github.bytebase/dbhub MCP server
3,362★ repoMinimal, token-efficient Database MCP Server for PostgreSQL, MySQL, SQL Server, SQLite, MariaDB
grafana/mcp-grafana/io.github.grafana/mcp-grafana MCP server
3,358★ repoAn MCP server giving access to Grafana dashboards, data and more.
giancarloerra/socraticode/io.github.giancarloerra/socraticode MCP server
3,253★ repoMCP server for local codebase indexing, semantic search, and code dependency graphs.
codeaashu/claude-code/io.github.divu2025/codemaster MCP server
no license3,221★ repoExplore the Claude Code CLI source — browse tools, commands, search code, and more.
Ar9av/obsidian-wiki/.skills/claude-history-ingest Skill
3,165★ repoFramework for AI agents to build and maintain a digital brain through Obsidian wiki using Karpathy's LLM Wiki pattern
apify/apify-mcp-server/com.apify/apify-mcp-server MCP server
3,160★ repoExtract data from any website with thousands of scrapers, crawlers, and automations on Apify Store ⚡
blazickjp/arxiv-mcp-server/io.github.blazickjp/arxiv-mcp-server MCP server
3,047★ repoSearch arXiv papers, download full text, semantic search, citation graphs, and alerts via MCP.
butterbase-ai/butterbase-oss/io.github.butterbase-ai/mcp MCP server
2,953★ repoButterbase MCP server — manage your backend: schemas, auth, functions, storage, RAG, deploys.