Claude skill · Development · Anthropic

Claude API skill by Anthropic

A large reference for building on the Claude API and SDKs: current model IDs, pricing, streaming, tool use, caching and agents, for seven languages plus curl.

Command: /claude-api
Licence: Apache-2.0, copying allowed with attribution
Folder: 96 files, 2.0 MB
Runs code: 4 scripts

Overview

What it does

This skill exists because a model’s training data goes stale. It tells Claude not to answer API questions from memory and gives it a current reference instead: model IDs, parameters, thinking and effort settings, prompt caching, the provider clients for Bedrock, Vertex AI and Foundry, and Managed Agents.

Language-specific guides for Python, TypeScript, Java, Go, Ruby, C#, PHP and curl sit in subfolders, so Claude loads only the one for your code. At over 600 lines, its SKILL.md is longer than Anthropic’s own recommended limit. For model IDs, our Claude model IDs page is kept in sync with the same docs.

Triggers

When Claude uses it

A prompt that names Claude, Anthropic, a Claude model or the Anthropic SDK.
Questions about LLM pricing, model choice, limits or caching.
LLM-shaped tasks where no other provider is in use; it stands aside when the prompt or the project uses another provider.

Its description, which Claude matches against

Reference for the Claude API / Anthropic SDK — model ids, pricing, params, streaming, tool use, MCP, agents, caching, token counting, model migration. TRIGGER — read BEFORE opening the target file; don't skip because it "looks like a one-liner" — whenever: the prompt names Claude/Anthropic in any form (Claude, Anthropic, Fable, Opus, Sonnet, Haiku, anthropic, @anthropic-ai, claude-*, us.anthropic.*, [1m]); the user asks about an LLM (pricing/model choice/limits/caching) — never answer from memory; OR the task is LLM-shaped with provider unstated (agent/MCP/tool-definition/multi-agent/RAG/LLM-judge/computer-use; generate/summarize/extract/classify/rewrite/converse over NL; debugging refusals/cutoffs/streaming/tool-calls/tokens). SKIP only when another provider is being worked on (overrides all triggers): OpenAI/GPT/Gemini/Llama/Mistral/Cohere/Ollama named in the query; OR grep -rE 'openai|langchain_openai|google.generativeai|genai|mistralai|cohere|ollama' over the project hits (run this grep FIRST if no provider named — don't Read the file).

What it needs

Nothing to install.
Its name contains “claude”, a word the Agent Skills naming rules reserve, so a claude.ai or Skills API upload is likely to be refused. Claude Code accepts it, and already ships with its own claude-api skill.

Install

How to install it

Claude Code

Plugin (type in Claude Code)

/plugin marketplace add anthropics/skills
/plugin install claude-api@anthropic-agent-skills

The claude-api plugin installs this skill only. Its skills are then namespaced, such as /claude-api:claude-api.

Or copy the folder (shell)

git clone --depth 1 https://github.com/anthropics/skills.git
mkdir -p ~/.claude/skills
cp -r skills/skills/claude-api ~/.claude/skills/

That makes it a personal skill, available as /claude-api in all your projects. Copy it into .claude/skills/ in a repository instead to share it with your team.

claude.ai and the Claude API

Zip the skill folder, then in claude.ai open Customize > Skills, choose + and Create skill, then Upload a skill. Code execution must be turned on.

Upload with the Python SDK

from anthropic import Anthropic
from anthropic.lib import files_from_dir

client = Anthropic()
skill = client.skills.create(files=files_from_dir("claude-api"))
print(skill.id)  # use as a custom skill_id with the code execution tool

The checks below found 2 problems that claude.ai and the API reject. Fix them before uploading.

Checks

SKILL.md validator results

Its SKILL.md (616 lines) run through our SKILL.md validator, with the folder’s file list, for both targets. These check the file’s format, not whether the skill works well.

Claude Code

0 errors, 5 warnings, 1 tip

warningNames can’t contain the reserved word “claude”. Claude Code accepts it, but claude.ai, the API and other agents don’t.
warningClaude Code already has /claude-api. Your skill replaces it (but not its aliases), which can be confusing. Pick a unique name.
warningThe description is 1,068 characters; the limit is 1,024. Claude Code loads it, but claude.ai and the API reject it.
warningSKILL.md has 608 lines of content. Anthropic recommends under 500: move detail into reference files that load only when needed.
warningClaude Code reads $0, $1 and so on as argument placeholders, so text like $10.00 (a price or a spreadsheet reference) can be replaced when the skill runs. Put a backslash before the $, such as \$5.00, to keep it as written.
tipSKILL.md doesn’t mention csharp/claude-api/batches.md, csharp/claude-api/files-api.md, csharp/claude-api/streaming.md, csharp/claude-api/tool-use.md, curl/examples.md and 72 more. Say what each file is for and when to read or run it, or Claude won’t know to use it.

claude.ai and the API

2 errors, 1 warning, 1 tip

errorNames can’t contain the reserved word “claude”.
errorThe description is 1,068 characters; the limit is 1,024.
warningSKILL.md has 608 lines of content. Anthropic recommends under 500: move detail into reference files that load only when needed.
tipSKILL.md doesn’t mention csharp/claude-api/batches.md, csharp/claude-api/files-api.md, csharp/claude-api/streaming.md, csharp/claude-api/tool-use.md, curl/examples.md and 72 more. Say what each file is for and when to read or run it, or Claude won’t know to use it.

Folder

What’s in the folder

LICENSE.txt11 KB
SKILL.md103 KB
csharp/5 files, 28 KB
curl/2 files, 20 KB
go/5 files, 42 KB
java/5 files, 44 KB
php/6 files, 34 KB
python/7 files, 105 KB
ruby/4 files, 19 KB
shared/54 files, 1.5 MB
typescript/6 files, 67 KB

SKILL.md sections

Building LLM-Powered Applications with Claude
Before You Start
Output Requirement
Defaults
Warning: API Drift - Your Training Prior May Be Stale
Subcommands
Language Detection
Language-Specific Feature Support
Which Surface Should I Use?
Building an Agent: Four Approaches
Should I Build an Agent?
Architecture
Current Models (cached: 2026-10-06)
Claude Fable 5.1 (claude-fable-5-1) - most capable widely released model
Claude Opus 5.5 (claude-opus-5-5) - the current Opus and the default model
Claude Sonnet 5.5 (claude-sonnet-5-5) - the current Sonnet: speed and capability for everyday coding, agent, and enterpr
Claude Haiku 5.5 (claude-haiku-5-5) - the current Haiku
Authentication (Quick Reference)
Thinking & Effort (Quick Reference)
Compaction (Quick Reference)
Prompt Caching (Quick Reference)
Fast Mode (Quick Reference)
Task Budgets (Quick Reference)
Provider Clients (Quick Reference)
Amazon Bedrock
Microsoft Foundry
Google Cloud Vertex AI
Context Editing (Quick Reference)
Mid-Conversation System Messages (Quick Reference)
Managed Agents (Beta)
Server Tools (Quick Reference)
Document & File Input (Quick Reference)
Tool Use Patterns (Quick Reference)
Other API Surfaces (Quick Reference)
Workload Identity Federation (Quick Reference)
Reading Guide
Quick Task Reference
When to Use WebFetch
Common Pitfalls

Source

Licence and source

Published by Anthropic in anthropics/skills/skills/claude-api, last changed 2026-10-09.
Licence: Apache-2.0, copying allowed with attribution. Our copy of the SKILL.md is unmodified, served with its licence.
Read from commit dbd4588; write-up checked 2026-10-14.
Before installing, read the SKILL.md on GitHub and its scripts. Anthropic advises using skills only from sources you trust (security considerations).

All Claude skills Write your own skill Slash command builder Guide: subagents, skills and commands

What it does

When Claude uses it

What it needs

How to install it

Claude Code

claude.ai and the Claude API

SKILL.md validator results

Claude Code

claude.ai and the API

What’s in the folder

SKILL.md sections

Licence and source

More skills