claude-api Skill
Skill Claude Code AI Agents & AI App DevelopmentClaude Code Customization & WorkflowThe explanation below is AI-generated. Please verify it against the sources.
This item is the "claude-api" skill, a reference document that guides an AI coding assistant on building applications with the Claude API and Anthropic SDKs. According to SKILL.md, it covers model IDs, context windows, and pricing; request parameters such as thinking/effort, task budgets, and prompt caching; streaming, tool use, Managed Agents, and server-side tools like web search and code execution; and guidance for migrating existing code to newer models. It also defines when the skill should be read (e.g., whenever Claude/Anthropic is named, or the task is LLM-shaped with the provider unstated) and when it should be skipped (e.g., when another provider such as OpenAI, Gemini, or Llama is already in use). SKILL.md instructs the assistant to first detect the project's programming language and then consult language-specific files (Python, TypeScript, Java, Go, Ruby, C#, PHP, cURL) rather than inferring SDK usage from memory, since several API shapes are noted as having changed over time.
Overview
This is not a SaaS product but a "skill" — a structured reference document intended to be read by an AI coding assistant so it can look up authoritative details about the Claude API and Anthropic SDKs instead of relying on possibly outdated training knowledge, per SKILL.md.
What you can do with claude-api
- Look up current Claude model IDs, context windows, and per-token pricing from the "Current Models" table
- Determine whether a task calls for a single API call, a code-orchestrated workflow, Claude API + tool use, or Managed Agents, using the decision tree in SKILL.md
- Check parameter usage for adaptive thinking/effort, task budgets, prompt caching, compaction, and context editing
- Find language-specific documentation for streaming, tool use, batches, and the Files API (Python, TypeScript, Java, Go, Ruby, C#, PHP, cURL)
- Reference provider-specific client setup for Amazon Bedrock, Google Vertex AI, and Microsoft Foundry
- Follow migration guidance (shared/model-migration.md) when moving code to a newer Claude model
- Review the authentication credential-resolution order and a list of common implementation pitfalls documented in SKILL.md
Sources
Original description (English)
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).
History of claude-api
- Skill Added claude-api