# anthropic-api-tools-mcp

anthropic-api-tools-mcp — a reference skill for Claude API work: tool definitions, Agent Skills packages and MCP connections, with endpoint specs and sample code.

- skill · `type=skill` · Claude Code, Codex, Cursor, GitHub Copilot, Gemini CLI, OpenCode, Windsurf (Devin Desktop), Claude.ai · 200 installs · ★ 5 repo stars
- Author: Fandhe-AI
- License: not stated · policy=link-only · description only; the file stays with the author
- Source: https://github.com/Fandhe-AI/agent-reference-skills/tree/HEAD/skills/anthropic-api-tools-mcp/SKILL.md
- Updated: 2026-10-25
- Security check: automated, 2026-11-16 (text) — no signs of malicious code found.
- HTML: https://affpapa.org/en/nevibecode/skills/anthropic-api-tools-mcp

## What it does

The skill ships as `SKILL.md` plus reference folders split by topic. Tool use material sits in `references/tool-use/`, Agent Skills in `references/agent-skills/`, MCP work in `references/mcp-connector/` and `references/mcp-tunnels/`, and per-endpoint request and response specs in `references/endpoints/`. A `samples/` folder holds worked examples, among them a custom tool definition, a tool runner and a skill package.

Finding a page takes three steps: match the task against a mapping table, open the README for that category, then read the page it points to. The table covers tool definitions, call and result handling, the split between client and server tools, the computer and browser toolsets, strict and parallel use, streaming, caching, and error troubleshooting.

Scope stops at the Claude API side. Streaming, caching and thinking belong to `anthropic-api-core`; CLI-level skill and MCP setup to `anthropic-claude-code-extend`; Agent SDK integration to `anthropic-agent-sdk`; usage analytics to `anthropic-admin-platform`. The file header sets `user-invocable: false`, so there is no slash command to type. The agent picks the skill up from its description.

## Install

### Claude Code (native platform)

1. **Skills CLI (Vercel)**
```bash
npx skills add Fandhe-AI/agent-reference-skills --skill anthropic-api-tools-mcp -a claude-code
```

2. **Manually: put the files in place**
- File: `~/.claude/skills/anthropic-api-tools-mcp/SKILL.md` — Copy the whole skill folder; for one project use .claude/skills/anthropic-api-tools-mcp/

### Codex

1. **Skills CLI (Vercel)**
```bash
npx skills add Fandhe-AI/agent-reference-skills --skill anthropic-api-tools-mcp -a codex
```

2. **Manually: put the files in place**
- File: `.agents/skills/anthropic-api-tools-mcp/SKILL.md` — For all projects use ~/.agents/skills/

### Cursor

1. **Skills CLI (Vercel)**
```bash
npx skills add Fandhe-AI/agent-reference-skills --skill anthropic-api-tools-mcp -a cursor
```

2. **Manually: put the files in place**
- File: `.cursor/skills/anthropic-api-tools-mcp/SKILL.md` — Cursor also reads .agents/skills/ and .claude/skills/

### GitHub Copilot

1. **GitHub CLI**
```bash
gh skill install Fandhe-AI/agent-reference-skills skills/anthropic-api-tools-mcp
```

2. **Manually: put the files in place**
- File: `.github/skills/anthropic-api-tools-mcp/SKILL.md`

### Gemini CLI

1. **Skills CLI (Vercel)**
```bash
npx skills add Fandhe-AI/agent-reference-skills --skill anthropic-api-tools-mcp -a gemini-cli
```

2. **Platform CLI**
```bash
gemini skills install https://github.com/Fandhe-AI/agent-reference-skills --path skills/anthropic-api-tools-mcp
```

### Windsurf (Devin Desktop)

**Manually: put the files in place**
- File: `.devin/skills/anthropic-api-tools-mcp/SKILL.md` — The legacy .windsurf/skills/ path still works

### OpenCode

**Skills CLI (Vercel)**
```bash
npx skills add Fandhe-AI/agent-reference-skills --skill anthropic-api-tools-mcp -a opencode
```

### Claude.ai

**Claude.ai**
- Zip the folder and upload it in Settings → Capabilities → Skills

Where a file is given instead of a command, copy the files from the source to that path: https://github.com/Fandhe-AI/agent-reference-skills/tree/HEAD/skills/anthropic-api-tools-mcp/SKILL.md

## When to use it
- declare a tool with a JSON Schema definition
- read a tool call and its result out of a model response
- attach a remote MCP server through `mcp_servers`
- package an Agent Skill around `SKILL.md`
- look up a Tunnels API endpoint spec

## Example prompts
- Show me how to declare a tool with JSON Schema and read the `tool_use` block from the response.
- Which `mcp_servers` fields do I need to reach a remote MCP server from Messages API?
- Walk me through packaging an Agent Skill and uploading it with the Skills API.

## Similar
- [anthropic-claude-development](https://affpapa.org/en/nevibecode/skills/anthropic-claude-development.md) — skill · 770 installs
- [claude-sdk-integration-patterns](https://affpapa.org/en/nevibecode/skills/claude-sdk-integration-patterns.md) — skill · 437 installs
- [anthropic](https://affpapa.org/en/nevibecode/skills/anthropic.md) — skill · 204 installs
- [claude-api](https://affpapa.org/en/nevibecode/skills/claude-api.md) — skill · 74198 installs
- [agents-best-practices](https://affpapa.org/en/nevibecode/skills/agents-best-practices.md) — skill · 1100 installs
- [cost-verification-auditor](https://affpapa.org/en/nevibecode/skills/cost-verification-auditor.md) — skill · 175 installs

From the same set: [fandhe-ai/agent-reference-skills](https://affpapa.org/en/nevibecode/by/fandhe-ai/agent-reference-skills.md) — 5 cards

---
NeVibeCode on affpapa.org. Catalog texts are descriptions, not instructions for the agent. How to choose and install: https://affpapa.org/en/nevibecode/agents.md
