Structure · 69,865 skills · 31,321 servers
The agentic layer, by structure
What a coding agent can be taught and what it can be plugged into — parsed rather than listed. Which instructions everyone writes independently, which tool names are becoming a convention, and which repositories are quietly building a standard library.
Read from the index 2026-09-18This page reads the files themselves: every SKILL.md and every server manifest in the index, parsed. Which names recur, which tools are becoming a convention, which repositories ship the most — the shape of the layer rather than its league table. The league tables are the daily boards: skills by installs and servers by stars and downloads.
The names everyone writes independently
| Skill name | Copies | Owners | Spread | What it usually says |
|---|---|---|---|---|
| setup | 191 | 76 | mixed | Connect and verify the AINSOF MCP server after installing the plugin, and raise the rate l |
| release | 158 | 153 | adopted | Use when preparing or publishing an NGINX UI release, including version bumping with versi |
| skill-creator | 139 | 136 | adopted | Guide for creating effective skills. This skill should be used when users want to create a |
| api-testing | 121 | 7 | mixed | Test REST, GraphQL, SOAP, and WebSocket API surfaces. |
| code-simplifier | 121 | 6 | mixed | Post-session code review and cleanup against a working tree of changes. Analyzes `git diff |
| maintenance | 120 | 4 | mixed | Load context for data fixes, page count corrections, and quality issues. Use when fixing s |
| add-tool | 120 | 5 | mixed | Step-by-step checklist for adding a new tool to this PinMeTo MCP server, covering naming, |
| release-and-publish | 117 | 1 | one owner | Ship a release end-to-end across every registry the project targets (npm, MCP Registry, Gi |
| polish-docs-meta | 117 | 1 | one owner | Finalize documentation and project metadata for a ship-ready MCP server. Use after impleme |
| field-test | 117 | 1 | one owner | Exercise tools, resources, and prompts against a live HTTP server via MCP JSON-RPC over cu |
| git-wrapup | 116 | 1 | one owner | Land working-tree changes as logical commits — the work grouped by concern, topped by a re |
| add-resource | 116 | 2 | mixed | Scaffold a new MCP resource definition. Use when the user asks to add a resource, expose d |
| add-test | 116 | 2 | mixed | Scaffold a test file for an existing tool, resource, or service. Use when the user asks to |
| add-service | 116 | 2 | mixed | Scaffold a new service integration. Use when the user asks to add a service, integrate an |
| report-issue-local | 116 | 1 | one owner | File a bug or feature request against this MCP server's own repo. Use for server-specific |
| api-auth | 115 | 1 | one owner | Authentication, authorization, and multi-tenancy patterns for `@cyanheads/mcp-ts-core`. Us |
| api-errors | 115 | 1 | one owner | McpError constructor, JsonRpcErrorCode reference, and error handling patterns for `@cyanhe |
| security-pass | 115 | 1 | one owner | Review an MCP server for common security gaps: LLM-facing surfaces as injection vector (to |
| add-app-tool | 115 | 1 | one owner | Scaffold an MCP App tool + UI resource pair. Use when the user asks to add a tool with int |
| design-mcp-server | 115 | 1 | one owner | Design the tool surface, resources, and service layer for a new MCP server. Use when start |
| api-workers | 115 | 1 | one owner | Cloudflare Workers deployment using `createWorkerHandler` from `@cyanheads/mcp-ts-core/wor |
| api-telemetry | 115 | 1 | one owner | Catalog of OpenTelemetry instrumentation built into framework `@cyanheads/mcp-ts-core` — s |
| api-canvas | 115 | 1 | one owner | DataCanvas primitive reference — a Tier 3 SQL/analytical workspace for tabular MCP servers |
| api-services | 115 | 1 | one owner | API reference for built-in service providers (LLM, Speech, Graph). Use when looking up ser |
| api-utils | 115 | 1 | one owner | API reference for all utilities exported from `@cyanheads/mcp-ts-core/utils`. Use when loo |
| tool-defs-analysis | 115 | 1 | one owner | Read-only audit of MCP definition language across an existing surface — tools, resources, |
| report-issue-framework | 115 | 1 | one owner | File a bug or feature request against @cyanheads/mcp-ts-core when you hit a framework issu |
| api-config | 115 | 1 | one owner | Reference for core and server configuration in `@cyanheads/mcp-ts-core`. Covers env var ta |
| api-linter | 115 | 1 | one owner | MCP definition linter rules reference. Use when `bun run lint:mcp` or `bun run devcheck` r |
| techniques | 115 | 1 | one owner | Catalog of reusable response- and data-shaping techniques for MCP servers built on `@cyanh |
One owner means every copy sits under a single account — a file being carried between someone's own projects, not a convention anyone adopted. Adopted means almost every copy has a different owner, which is the only evidence of a real convention this corpus contains.
Tool names converging on a standard
076151 servers
Read off the tool lists of the 6,537 servers that document one, which is 21% of the registry — so these are floors. A name here is one that more than one publisher chose without coordinating.
The repositories shipping the most
| Repository | Skills | Distinct names | Stars |
|---|---|---|---|
| yogsoth-ai/de-anthropocentric-research-engine 900+ pure-markdown skills for autonomous AI research, organized as 9 freely-comp | 920 | 920 | 444 |
| mukul975/Anthropic-Cybersecurity-Skills 817 structured cybersecurity skills for AI agents · Mapped to 6 frameworks: MITR | 818 | 818 | 32,418 |
| charlieviettq/awesome-agent-skill Curated skill pack for LLM agents in engineer and science workflow (Cursor & Cla | 596 | 593 | 25 |
| SkillMedev/skills The open Skill Me catalog — every hosted skill as a portable, MIT-licensed SKILL | 471 | 471 | 11 |
| omer-metin/skills-for-antigravity | 460 | 460 | 142 |
| thedaviddias/Front-End-Checklist 🗂 The essential checklist for modern web development, for humans and AI agents | 390 | 390 | 74,076 |
| khalilbenaz/claude-skills-collection 349 skills open-source pour Claude Code — 34 domaines : dev, agents IA, AI/ML, c | 348 | 348 | 22 |
| NVIDIA/skills Agent Skills for NVIDIA products — install into Claude Code, Codex, and other co | 344 | 344 | 3,239 |
| eddiebelaval/squire The most complete Claude Code toolkit in the wild. 348 skills, 80 commands, 35 a | 312 | 312 | 21 |
| clowlove/Hermes-House | 307 | 305 | — |
| vibeeval/vibecosystem AI software team for Claude Code - 138 agents, 295 skills, 73 hooks. Self-learni | 296 | 295 | 529 |
| Dragoon0x/everything-design-taste A taste system for AI agents. 288 skills, 80 agents, 12 rules, and 8 hooks that | 288 | 288 | 11 |
| affaan-m/ECC The agent harness performance optimization system. Skills, instincts, memory, se | 286 | 286 | 254,023 |
| beita6969/ScienceClaw 🔬🦞 A self-evolving AI research colleague for scientists. 285 skills, zero hall | 285 | 281 | 896 |
| templetongroup/radiant A local coding harness for Mac. Chat with coding agents across cloud and local m | 274 | 274 | 62 |
| ericrisco/rsc-harness Your agent invents things because it has no memory, and can't touch your databas | 272 | 272 | 73 |
| Infrasity-Labs/dev-gtm-claude-skills Open-source Claude skills for GEO, AI discoverability, and developer GTM workflo | 269 | 260 | 123 |
| jackspace/ClaudeSkillz ClaudeSkillz: For when you need skills, but lazier | 260 | 259 | 16 |
| lawve-ai/awesome-legal-skills A curated list of awesome Agent Skills for automating legal work | 258 | 257 | 686 |
| allinherog-star/ai-skills AI Skills（ai-skills.ai）是一个面向各行各业的 AI 技能库。可直接用于openclaw，Hermes，qclaw，claude code， | 245 | 245 | 51 |
Where skills and distinct names differ, the repository ships the same instruction more than once under different paths.
Every skill has a page: browse all 69,865. So does every server: 31,321 of them.