CCA-F: Claude Certified Architect — Foundations

By Anthropic · Question Mode

QuestionQ54

Context Management & Reliability

You are building developer productivity tools with the Claude Agent SDK. The agent assists engineers in exploring unfamiliar codebases, understanding legacy systems, generating boilerplate code, and automating repetitive tasks. It relies on the built-in tools (Read, Write, Bash, Grep, Glob) and integrates with Model Context Protocol (MCP) servers.

Suppose your agent has already analyzed a complex service module by:

  • Reading 23 source files
  • Tracing request flows
  • Identifying error handling patterns

A developer now wants to compare two different testing strategies before committing to one:

  1. End-to-end tests with mocked external services
  2. Snapshot tests capturing expected outputs

They need to develop both approaches independently so they can evaluate the trade-offs between them.

How should you manage the sessions in this case?

Explanation

Session forking in the Claude Agent SDK (using resume with fork_session=True, or the fork_session() helper) creates a new session that starts with the complete message history of the parent analysis session but then diverges independently from that point forward. This lets you preserve and reuse the costly upfront investigation (reading the 23 files, tracing flows, identifying error patterns) while creating two isolated branches — one for each testing strategy — so the developer can evaluate both approaches in parallel without either branch's exploration or decisions influencing the other, and without re-running the analysis or manually re-establishing context via exported notes.

Learn more

Community Discussion

No comments yet. Be the first to start the discussion!