anthropic-api

Purpose

This skill integrates with Anthropic's Claude API to leverage advanced AI models for tasks requiring tool use and extended context, enabling complex interactions like function calling and long-form processing.

When to Use

Use this skill for scenarios involving sophisticated AI reasoning, such as analyzing large documents, generating code with tools, or handling multi-step workflows. It's ideal when standard models fall short, like in research, automation, or interactive applications.

Key Capabilities

• Access Claude models (e.g., claude-3-haiku, claude-3-sonnet) via API for text generation and completion.
• Support tool use, allowing the AI to call external functions based on JSON-defined tools.
• Handle extended contexts up to 200K tokens for processing long inputs.
• Real-time streaming responses for interactive sessions.
• Fine-grained control over parameters like temperature (0.0-1.0) and top_p (0-1).

Usage Patterns

To use this skill in OpenClaw, first set the API key as an environment variable (e.g., export ANTHROPIC_API_KEY=your_key). Invoke the skill via OpenClaw's execute command, passing parameters like model and prompt. For HTTP-based integration, make POST requests to the API endpoint. Always include authentication headers. Structure requests with JSON bodies containing "messages" array for conversation history and "tools" for function definitions.