Claude Agent SDK
Anthropic's framework for building agents on the same loop, tools and context management that power Claude Code.
In more detail
It handles the loop, built-in tools, sessions and compaction. A canUseTool callback lets your code approve, deny or edit each tool call before it runs.
Related terms
- agent SDK — A library that packages the agent loop with tools, budgets, saved state, tracing and hooks for human approval.
- human approval — Pausing an agent before a risky action so a person can approve, edit or reject it, then resuming from saved state.
- Claude Code — Anthropic's agentic coding tool, which reads a codebase, edits files and runs commands from the terminal or an IDE.
Taught in
- Agent SDKs And Human Approval — AI Engineering: LLM Apps, RAG & Agents
References
- Claude Agent SDK: How the agent loop works
- Claude Agent SDK: Handle approvals and user input (canUseTool)