Playbook

The Claude Code Playbook

Practical, job-shaped guides for running Claude Code and other coding agents on a shared codebase — the loops, hooks, skills, and MCP setups agent teams actually use. Each page shows the workflow done end to end, working on top of First-Tree's shared context, not just a definition.

The Claude Code Loop

The plan-act-verify loop agents run all day, and how to keep it on the rails.

Read →

Claude Code MCP

Wire Model Context Protocol servers into Claude Code so agents reach your tools and data.

Read →

Claude Code Hooks

Run your own scripts at each step of the agent loop to enforce checks and guardrails.

Read →

Claude Code Skills

Package repeatable workflows as skills every agent on the team can invoke.

Read →

Claude Code Commands

Build custom slash commands that turn multi-step jobs into one reliable call.

Read →

Claude Code Best Practices

The habits that keep Claude Code fast, safe, and predictable on a real codebase.

Read →

CLAUDE.md

Write the CLAUDE.md project file that gives every agent the same starting context.

Read →

AGENTS.md

The open, cross-tool standard agents read — what goes in it and how it relates to CLAUDE.md.

Read →

AGENTS.md vs CLAUDE.md

Which agent context file to use, when, and how to keep them from drifting on a team.

Read →

Agentic Coding

How agentic coding actually works when a team of agents shares one repo.

Read →

AI Agent Workflows

Design end-to-end workflows that hand work cleanly between agents and people.

Read →

AI Code Review

Use agents to review diffs without losing the human judgment that matters.

Read →

Orchestrating Coding Agents

Route work to parallel agents — and give them the shared context to agree.

Read →

AI Agent Teams (the pillar)

The big picture: how shared context lets a whole team of agents work on one codebase.

Read →