Claude Agent SDK

Patterns this framework supports natively

• ★★ ReAct — Interleave a single thought, a single tool call, and a single observation per step so the agent reasons over fresh evidence.
• ★★ Tool Use — Let the LLM produce typed calls against an external toolkit instead of producing free-form text the surrounding system has to parse.
• ★ Computer Use — Let the model drive a desktop end-to-end via screenshots plus virtual mouse/keyboard tool calls instead of bespoke per-app APIs.
• ★★ Code Execution — Let the model emit code, run it in a sandbox, and treat the run as the answer instead of trusting the model to compute in its head.
• ★ Agent Skills — Package author-time procedures (markdown + optional resources) the agent loads on demand for specific task types.
• ★★ Step Budget — Cap the number of tool calls or loop iterations the agent is allowed within a single request.
• ★ Subagent Isolation — Run subagents in isolated workspaces so their writes do not collide and parallelism is safe.
• ★★ Approval Queue — Queue agent-proposed actions for asynchronous human review while the agent continues other work.