Codex Agent Skills
Reusable workflows for Codex CLI, IDE, and app.
Codex skills are SKILL.md directories with instructions, optional scripts, references, assets, and agent metadata. They can be distributed through plugins.
$skill-creator
The command center for MCPs, skills, plugins, tools, docs, CLIs, and frameworks across Codex, Grok Build, Cursor, Claude, Gemini, GitHub Copilot, Ollama, Vercel, and Supabase.
10
Providers
24
Resources
8
Resource types
3
Guides
Ready for Supabase catalog data
Public content is database-backed with RLS; signed-in users get a private resource vault.
Explore
Reusable workflows for Codex CLI, IDE, and app.
Codex skills are SKILL.md directories with instructions, optional scripts, references, assets, and agent metadata. They can be distributed through plugins.
$skill-creatorCode-first agent orchestration in TypeScript and Python.
The Agents SDK is for applications that own orchestration, tool execution, approvals, state, handoffs, tracing, and evaluation.
npm install @openai/agents zodUse remote MCP tools through the Responses API.
OpenAI Responses API can import tools from remote MCP servers that support Streamable HTTP or HTTP/SSE, then call those tools as part of a model run.
SKILL.md workflows for Claude Code and Claude API.
Claude Code skills live in personal, project, plugin, or enterprise locations and can include scripts, templates, examples, and references.
mkdir -p ~/.claude/skills/summarize-changesBundle skills, agents, hooks, MCP servers, and monitors.
Claude Code plugins are self-contained directories that extend Claude Code with skills, agents, hooks, MCP servers, LSP servers, and monitors.
Connect Claude Messages API to remote MCP servers.
Anthropic MCP connector lets Claude API requests connect directly to remote MCP servers using a beta header and remote HTTP/SSE tools.
Terminal coding agent from xAI.
Grok Build is an early beta coding agent and CLI with interactive TUI sessions, plan/review/approve flow, headless mode, ACP, AGENTS.md, plugins, hooks, skills, MCP servers, and subagents.
curl -fsSL https://x.ai/cli/install.sh | bashxAI coding model for agentic harnesses.
grok-build-0.1 is xAI's coding model for agentic coding tasks including web development, debugging, and MCP support.
Connect Cursor agents to tools and data sources.
Cursor supports Model Context Protocol for connecting external tools and data to its agent workflows.
Providers
OpenAI
Agent platform
Codex, Agents SDK, Responses API, skills, plugins, and MCP integrations for agentic software work.
DocsAnthropic Claude
Agent platform
Claude Code, Agent Skills, MCP connector, and plugin-style customization for Claude workflows.
DocsxAI Grok
Agentic coding
Grok Build CLI and grok-build-0.1 model for terminal coding, headless scripting, ACP, skills, plugins, hooks, and MCP.
DocsCursor
AI IDE
Agent mode, rules, MCP configuration, background agents, and coding workflow customization.
DocsGoogle Gemini
Agent platform
Gemini CLI, Gemini Code Assist agent mode, Gemini Docs MCP, API skills, and Google ADK.
DocsGitHub Copilot
Developer agent
Copilot cloud agent, repository MCP configuration, GitHub MCP server, and custom agents.
DocsOllama
Local models
Local model runtime with tool calling, structured outputs, OpenAI compatibility, and agent-loop friendly APIs.
DocsVercel
AI app platform
AI SDK agents, MCP tools, Vercel MCP, deployment workflows, and production infrastructure for agent apps.
DocsSupabase
Backend platform
Postgres, Auth, Edge Functions, MCP tooling, AI prompts, and agent-ready database workflows.
DocsModel Context Protocol
Open standard
Open protocol, registry, SDKs, inspector, server metadata, and cross-client integration patterns.
DocsSource ledger
OpenAI Developers
Skill structure, progressive disclosure, locations, and plugin distribution guidance.
SourceOpenAI Developers
Current install commands for TypeScript and Python Agents SDK.
SourceSupabase Docs
OAuth callback URL format and GitHub provider configuration.
SourcexAI Docs
Install, interactive TUI, headless mode, config, skills/plugins/hooks/MCP discovery.
SourceGoogle AI for Developers
Gemini Docs MCP endpoint, skills.sh commands, and verification steps.
SourceMicrosoft Playwright
Install snippets, supported clients, capabilities, and security note for code execution.
SourceVercel Docs
Official Vercel MCP endpoint, supported clients, setup commands, and security guidance.
SourceModel Context Protocol
Preview registry, metadata, namespace verification, discovery model, and security notes.
Source