Practice mode · free · unlimited · not certifiable

Multi-Agent Orchestration and Naming

Unscored tasks with hints and instant feedback. Practice results never become credentials.

Short lesson

Working with agents safely · 10 min

Name each agent by role and task, isolate them in worktrees, hand off through files, review every diff, never paste secrets. Tool notes (2026-10): Claude Code (CLAUDE.md, agents folder), Codex CLI and others (AGENTS.md), Gemini CLI (GEMINI.md), Cursor, Copilot — verify names in each tool's docs.

L1Name your agentterminal

Start one mock agent for the README task using a role-and-task name (e.g. builder-readme), in ~/shop, and give it the task “update readme”.

L3Two agents, two worktreesterminal

In ~/shop create worktrees ../shop-builder-login (branch agent/builder-login) and ../shop-tester-login (branch agent/tester-login). Start builder-login (profile steady) in the first to implement the “login form” and tester-login (profile steady) in the second to write “tests”. Approve good work, write a hand-off note per agent in ~/shop/handoff/<agent>.md, and commit each worktree with a message that names the agent.

L4Merge one at a timeterminal

Branches agent/builder-login and agent/tester-login are ready in ~/shop. Merge them into main one at a time, running `npm test` (simulated: sh test.sh) after each merge, and record each step in ~/shop/RUNLOG.md.