Coding Workflows

This tutorial demonstrates server-side prompt versioning and rollback for Managed Agents, enabling PMs to update agent prompts without code deployments. It covers creating an agent (v1), evaluating it against a labeled test set, shipping an updated prompt (v2), detecting performance regressions, and rolling back by pinning sessions to a specific version. The workflow replaces traditional code-based prompt management with immutable versioned prompts that can be quickly reverted if issues arise.

This tutorial demonstrates building a webhook-triggered SRE incident response agent using Claude Managed Agents that automatically investigates production alerts, consults runbooks, proposes infrastructure fixes via pull requests, and gates merging behind human approval. The agent combines built-in sandbox tools (bash, read, edit) with custom tools for PR management and human-in-the-loop approval, providing complete audit trails in the Anthropic Console. The example uses mocked PagerDuty, GitHub, and Datadog integrations to focus on agent patterns, with guidance for swapping in real services.

This tutorial introduces the Claude Managed Agents API by walking through a practical debugging workflow where an agent iteratively fixes failing tests in a Python package. It covers the three core resources (Agent, Environment, Session), file mounting, and the streaming event loop pattern. The example demonstrates how agents autonomously discover the iterate-observe-fix loop by running tests, reading failures, editing code, and repeating until all assertions pass.

The Claude Agent SDK provides built-in session management functions to build conversation history sidebars without writing custom parsers. This cookbook demonstrates how to list, read, rename, tag, and fork sessions stored as JSONL transcripts on disk. By leveraging SDK primitives like list_sessions(), get_session_messages(), and rename_session(), developers can implement session browsers for desktop apps, IDE extensions, and chatbots without managing file I/O directly.

Claude Code v2.1.91 introduces MCP tool result persistence overrides supporting up to 500K characters, adds shell execution disabling options, enables multi-line prompts in deep links, and allows plugins to ship executables. The release includes critical fixes for transcript chain breaks, terminal key handling, plan mode persistence in remote sessions, and JSON schema validation, plus performance improvements and enhanced API guidance.