Explore AI Agent Skills & Claude Prompts
Discover open-source agent skills for Claude Code, Codex, ChatGPT, and any tool that uses SKILL.md.
Enter through keywords, occupations, creators, and GitHub sources to see what kinds of skills are emerging across domains.
Use the same catalog through the API
Connect 381,784 public skills to your own search, analytics, or agent workflow with the REST API.
Querying local SQLite index...
airtable-connector
by selrai-companyConnect and operate Airtable via the official first-party Airtable MCP server (https://mcp.airtable.com/mcp). Drives the entire setup autonomously through airtable.com/create/tokens in a Playwright MCP browser: clicks Create new token, fills the name, ticks the four required scope checkboxes, selects All current and future bases, clicks Create token, reads the Personal Access Token from the DOM, and registers the MCP server with the token as a Bearer header. The only human moment is signing in to Airtable once. Use this skill when the user asks to set up Airtable, connect their bases, list tables, read or create records, or update their database schema. On first use run Phase 1 to configure the MCP server and authenticate before attempting tool calls.
aws-connector
by selrai-companyInstall and operate the AWS connector autonomously. Drives console.aws.amazon.com end-to-end inside a Playwright MCP browser: navigates to IAM, finds or creates a least-privilege IAM user named 'claude-assistant' with the right managed policy attached (ReadOnlyAccess for read-only access, PowerUserAccess for read-and-write — never AdministratorAccess), creates an access key for that user, DOM-extracts the access key ID and secret access key from the post-create modal (AWS shows the secret once and never again), and writes ~/.aws/credentials + ~/.aws/config autonomously. The user's only manual moments are signing in to the AWS Console once and approving any MFA prompt their account requires. Never mints root access keys — verifies post-install via sts:GetCallerIdentity that the keys belong to an IAM user (arn:aws:iam::*:user/*), not the root account (arn:aws:iam::*:root), and refuses to ship root credentials. Read and operate AWS services via the aws CLI (S3, EC2, Lambda, DynamoDB, IAM read-side, Cost Explorer
docusign-connector
by selrai-companyConnect and operate DocuSign eSignature on the user's behalf via the official DocuSign eSignature REST API at `demo.docusign.net/restapi/v2.1` (sandbox) or `<base_uri>/restapi/v2.1` (production). Phase 1 is a 7-step Playwright-driven install: drive `admindemo.docusign.com/api-integrator-key` to create an Integration Key, DOM-extract the key, generate + capture a Secret Key from the one-time-reveal modal, register `http://localhost:8080/callback` as the Redirect URI, save, then drive DocuSign's OAuth Authorization Code Grant + PKCE flow with a local-listener exchange — the user signs in once, Claude auto-clicks Allow on the consent screen, the listener captures the code and exchanges it for an access token + 30-day refresh token in sub-second latency, and writes both into `~/.claude.json`. No DocuSign hosted MCP server is involved — DocuSign's `mcp.docusign.com` gateway rejects the OAuth-issued tokens (empirical finding 2026-05-05, issue #213). Phase 2 calls DocuSign's REST API directly via Bash + `curl`, with
github-connector
by selrai-companyConnect and operate GitHub via the official GitHub remote MCP server (https://api.githubcopilot.com/mcp). Drives the entire setup autonomously through github.com/settings/personal-access-tokens/new in a Playwright MCP browser: fills the token name, picks expiration, ticks Repository permissions to match the user's read-only or read-and-write choice, clicks Generate token, reads the fine-grained Personal Access Token from the DOM, and registers the MCP server with the token as a Bearer header. The only human moments are signing in to GitHub once and any 2FA challenge their account requires. Use this skill when the user asks to set up GitHub, connect their GitHub account, or interact with repositories, issues, pull requests, commits, branches, releases, Actions workflows, or code search. On first use run Phase 1 to configure the MCP server and authenticate before attempting tool calls.
google-ads-connector
by selrai-companyConnect and operate Google Ads (read campaigns, ad groups, ads, keywords, search terms, audiences; pull reports for spend/conversions/ROAS; pause/resume campaigns and adjust budgets). Phase 1 captures credentials and Phase 1F wraps Google's official MCP server (github.com/googleads/google-ads-mcp) for read operations — Phase 2 reads call mcp__google_ads__search / list_accessible_customers / get_resource_metadata. Writes (pause/resume campaign, change budget) stay on direct REST because the official MCP is read-only. Each participant has their own Google Cloud OAuth client and their own developer token. Phase 0 picks between Test mode (immediate dev-token, only works with Google Ads test accounts) and Basic Access mode (live data, real customer, but Google reviews the dev-token application over ~1-3 business days). Phase 1 drives Google Cloud Console + Google Ads API Center + Google's OAuth consent flow + the optional 'New Google Ads account' wizard, all inside a Playwright MCP browser. Credentials persist at
hubspot-connector
by selrai-companyConnect and operate HubSpot CRM via the official @hubspot/mcp-server. Drives the entire setup autonomously through app.hubspot.com/private-apps/ in a Playwright MCP browser: opens the Create-a-private-app form, fills the name, walks the Scopes tab to tick the required CRM scopes, clicks Create app, and reads the Private App access token directly from the DOM. The only human moment is the user signing in to HubSpot once. Use this skill when the user asks to set up HubSpot, connect their CRM, or interact with contacts, companies, deals, tickets, notes, tasks, associations, properties, custom objects, or workflows. On first use, run Phase 1 to install and authenticate the connector before attempting any tool calls.
imessage-connector
by selrai-companyConnect the user's iMessage to Claude Code (Mac only) so they can text their assistant from any Apple device. Handles the Full Disk Access grant, plugin install, and self-chat test conversationally. Use this skill when the user says 'set up iMessage', 'connect my iMessage', 'text Claude from my iPhone', or asks about iMessage allowlists, Full Disk Access, or AppleScript replies.
klaviyo-connector
by selrai-companyConnect and operate Klaviyo (e-commerce-focused email + SMS marketing) — read profiles/subscribers, lists, segments, campaigns, flows (automations), campaign performance (opens/clicks/revenue); add profiles to lists, send draft campaigns, suppress profiles. Tier-1 connector for SMBs running e-commerce email + SMS, especially Shopify/WooCommerce stores. Direct-REST against Klaviyo API at https://a.klaviyo.com/api/ with `Authorization: Klaviyo-API-Key <key>` header and a required `revision` header pinning the API version (this SKILL uses `revision: 2025-10-15`). Single-mode connector — Klaviyo's free tier IS real data (up to 250 profiles, 500 email sends/month), no test vs live distinction. Phase 1 drives www.klaviyo.com via Playwright: signs participant in, navigates directly to the canonical `/settings/account/api-keys` URL, clicks Create Private API Key (which navigates to a full-page form at `/create-private-api-key`), names it `Claude Workshop Connector` with the Full Access Key access level, captures the
meta-business-suite-connector
by selrai-companyConnect the user's Instagram Business and Threads to Claude Code so Claude can read, publish, and manage content on their behalf via the @mikusnuz/meta-mcp server. Drives the entire setup autonomously through the Meta Developer Portal + Graph API Explorer in a Playwright MCP browser: creates the Meta Developer App, captures the App ID and App secret from the DOM, drives Graph API Explorer to mint the access token, exchanges short-lived for long-lived via curl, and writes ~/.claude.json directly. The only human moments are the user logging in to Facebook and ticking permission boxes in Meta's OAuth dialog. Phase 2 covers Instagram publishing (photos, videos, carousels, Reels, stories), media management, comments and replies, profile and account insights, hashtag search, mentions, DMs; Threads publishing (text, image, video, carousels, polls, link attachments), replies, search, profile, post insights; plus Meta platform token management. Use this skill when the user says 'connect my Instagram', 'install the Met
notion-pit-setup
by selrai-companyConnect the user's Notion workspace to Claude Code via a Personal Integration Token (PIT), so the resulting MCP entry is replicable to a headless agent server. Install is autonomous Playwright-driven: drive `notion.so/profile/integrations` in the Playwright MCP browser, detect login state and prompt sign-in only if needed, autonomously create a new internal integration (name + workspace + capabilities + submit), reveal and extract the PIT from the DOM via clipboard transit (the raw token NEVER appears in chat or tool-call returns), then register the MCP server entry via `claude mcp add --transport http --header`. The user's only manual moments are signing in to Notion (if not already), confirming the workspace + capabilities on the creation form, and choosing the page-access strategy (workspace teamspace connection if admin/Plus, otherwise per-page sharing). Use this skill when the user asks to set up Notion FOR THE AGENT SERVER, mentions needing Notion in a headless install, or explicitly asks for the PIT pa
pandadoc-connector
by selrai-companyConnect and operate PandaDoc via the official first-party PandaDoc MCP server (https://mcp.pandadoc.com/v1/mcp). Phase 1 is a 6-step Playwright-driven install: register the server with `claude mcp add`, open Claude Code's OAuth start URL inside the Playwright MCP browser, detect login state and prompt sign-in only if needed, auto-click 'Allow access' and 'Authorize' on the consent screen, auto-detect the callback via `browser_wait_for`, then verify with a read-only `mcp__pandadoc__*` smoke call. The user's only manual moment is signing in to PandaDoc inside the Playwright window. Use this skill when the user asks to set up PandaDoc, connect their account, create contracts or proposals from templates, send documents for signature, track signature status, find documents awaiting signature, search or filter documents, summarize or extract document text, expire or void documents, or any e-signature / document-automation workflow.
paypal-connector
by selrai-companyConnect and operate PayPal via the official @paypal/mcp server. Drives the entire setup autonomously through developer.paypal.com/dashboard/applications in a Playwright MCP browser: switches to the chosen Live or Sandbox tab, clicks Create App, fills the app name, captures Client ID and Secret from the DOM after clicking Show, silently exchanges them for an access token, and writes ~/.claude.json. The only human moments are the user signing in to PayPal once and answering live-vs-sandbox. Use this skill when the user asks to set up PayPal, connect their account, or interact with invoices, payments, orders, refunds, disputes, subscriptions, products, shipment tracking, or transactions. On first use, run Phase 1 to install and authenticate the connector before attempting any tool calls.
Browse Agent Skills by Occupation
23 major groups · 867 SOC occupations
Browse by Category
Explore agent skills organized by their primary use case
Explore the agent skills ecosystem by occupation and creator
SkillMD is not just a keyword search box. It is an open map that organizes public skills by occupation, creator, and repository, helping you see which workflows, judgment criteria, and domain habits people are writing for AI agents.
Then follow creators and GitHub repositories back to the source: compare the skills a team maintains, whether the repo is active, and how the README frames the work before you open, install, or reuse anything.
Use it three ways: learn an unfamiliar field by occupation, study how creators organize skills, then use source context to decide what is worth opening or reusing.
01 Map a field
Browse 23 occupation groups and 867 SOC roles to learn what skills exist in adjacent domains and how they break down real work.
02 Follow creators
Use creator and repository pages to inspect maintained skill collections, recent updates, and source context before trusting a result.
03 Search with sources
Search 1.7M+ collected skills, then use occupation tags, creators, and GitHub source context to decide what is worth opening.
Start with the occupation map, then follow creators and repositories back to real code. SkillMD helps explain why a skill is worth opening, not only what it is named.
Standardizing Agent Capabilities with SKILL.md and Model Context Protocol (MCP)
In the rapidly evolving landscape of artificial intelligence, LLM agents (Large Language Model agents) have transitioned from simple text predictors to autonomous problem solvers. To orchestrate complex, multi-step agentic workflows, developers require a standardized format to specify agent capabilities, prompt instructions, system rules, and database bindings. This is where SKILL.md and the Model Context Protocol (MCP) have emerged as standard developer paradigms. SkillMD serves as the central directory for indexing, exploring, and sharing these critical agent configurations.
Our open-source registry currently tracks over 1.7 million collected SKILL.md configurations and system prompts. By compiling agent configurations from active developers on GitHub, we bridge the gap between prompt engineering research and production execution. Whether you are building agents with Anthropic's Claude Code, OpenAI's GPT-4, Google's Gemini, or local models using Ollama and LlamaIndex, standardized skill definitions ensure your agents behave predictably across different runtime environments.
What is the Model Context Protocol (MCP)?
The Model Context Protocol (MCP) is an open-source standard designed to connect LLMs to data sources, developer tools, and external environments. MCP establishes a bidirectional communication channel between client applications (like Cursor, Claude Desktop, or custom agent systems) and servers hosting data or capabilities. Standardizing instructions via SKILL.md enables LLMs to query databases, read local files, execute terminal commands, and integrate third-party APIs. SkillMD allows you to find ready-to-run MCP servers and prompt instructions for various occupations and technical tasks.
The Structure of a Professional SKILL.md File
A valid SKILL.md configuration is designed to be easily read by humans and parsed by LLMs. It contains precise system instructions, trigger conditions, required parameters, and execution examples. Below is the typical architectural blueprint of a professional agent skill:
- Metadata & Core Scope: Declares the name of the skill, author details, target models, and a description of the capability.
- Triggers & Intent Detection: Details semantic triggers that help the agent decide when to invoke this skill.
- System Prompts: Explicit system-level instructions that direct the agent's behavior, personality, safety guardrails, and formatting preferences.
- Capabilities & Tools: Lists the files, databases, or APIs the agent must access to complete the tasks.
- Few-Shot Examples: Demonstrates real inputs and outputs, helping the model generalize behavior through in-context learning.
Optimizing Agent Workflows for Modern LLMs
Writing effective agent skills requires deep knowledge of prompt engineering. With the release of advanced reasoning models like Claude 3.5 Sonnet, ChatGPT o1, and DeepSeek-V3, prompt templates must focus on structured thinking. Developers are encouraged to use XML tags (e.g., <thought>, <context>, and <rules>) to isolate execution boundaries. Standardized prompts prevent agents from suffering from context drift, ensuring that long-running tasks remain aligned with the initial system parameters.
Exploring by SOC Occupations and Creator Profiles
What makes SkillMD unique is its taxonomy. Instead of simple text search, we parse and organize files according to the Standard Occupational Classification (SOC) system. This means you can discover skills written for Computer and Mathematical roles, Business and Financial operations, Legal, Design, and and Educational Instruction fields. By tracking creator profiles, developers can study how different teams organize their custom instructions, compare version updates, and fork public configs for specialized enterprise use cases.
SkillMD operates as a high-performance index running on a fast Go backend and a highly responsive Astro SSR frontend. All search queries execute in milliseconds, featuring smart debouncing to prevent multiple API requests while keeping user data secure. Join our community of developers to standardize your AI agent instructions and optimize your LLM prompting workflows today.
Frequently Asked Questions
A practical guide to agent skills: what they are, how to inspect them, and how SkillMD helps you explore the ecosystem.