← HOME

PROJECT

Claude Code Setup

ACTIVE

Power user environment for AI-assisted development

Claude CodeTypeScriptBashGitHub MCPContext7 MCP

ABOUT

Claude Code Setup is the personal development environment for AI-assisted engineering at Understory Labs. It's not a project with a URL — it's the infrastructure layer that everything else runs on top of.

The core is a library of 25+ global skills: reusable Claude Code commands that codify workflow patterns (/plan, /brainstorm, /commit, /log, /debug, /review, and more). Combined with MCP servers (GitHub integration, Context7 documentation lookup) and hooks (git event handlers, environment protection), it creates an environment where common operations are single commands.

The opusplan model profile is a key architectural detail: Opus during planning and architecture sessions, Sonnet during execution — so the most capable model is used where reasoning depth matters, and the faster model handles implementation. This setup is the scaffold for every other project in Understory Labs.

FEATURES

SHIPPED

  • 25+ global skills — /plan, /brainstorm, /commit, /log, and more

    Reusable Claude Code commands codifying common engineering workflow patterns

    MAR 2026
  • MCP servers — GitHub and Context7

    GitHub integration for PR/issue management; Context7 for live documentation lookup

    MAR 2026
  • Hooks — git and environment event handlers

    Pre/post tool hooks: .env protection, git integration, session events

    MAR 2026
  • opusplan model profile — tiered model selection

    Opus for planning sessions, Sonnet for execution — configured in settings.json

    MAR 2026

CHANGELOG

CLAUDE-CODEinfrastructuretooling

Power user setup — Steps 1 through 8

Features

  • Full Claude Code power user environment built across 8 steps
  • 20+ skills installed globally — /start, /plan, /brainstorm, /commit, /save, /tdd, and more
  • MCP servers configured: GitHub MCP for repo access, Context7 for live documentation
  • Hooks configured for pre/post tool events
  • CLAUDE.md updated with full work context, stack, and learning preferences

Infrastructure

  • Skills stored in ~/.claude/commands/ — available in all projects
  • toolkit.md created as single-source inventory of all capabilities
  • power-user-setup.md documents the step-by-step build history
  • opusplan model profile set as default — Opus during planning, Sonnet during execution

Lessons

  • The skill system makes Claude Code dramatically more consistent across sessions
  • CLAUDE.md is more valuable the more specific it is — vague instructions get ignored
  • Model selection matters: Opus for architecture decisions, Sonnet for building, Haiku for quick lookups
  • /save after every session is the habit that makes everything else work
CLAUDE-CODESHIPPEDinfrastructuretooling

Claude Code Setup — power user environment shipped

Features

  • Claude Code power user environment ships — 20+ global skills, MCP servers, hooks, and model profile operational
  • 20+ global skills installed: /plan, /brainstorm, /commit, /save, /tdd, /debug, /review, /start, /log, and more — common workflow patterns encoded as single commands
  • GitHub MCP + Context7 MCP configured — repo access and live documentation lookup available in any conversation
  • Hooks registered for git and file events — auto-format, lint, env protection, and toast notifications running silently in the background
  • opusplan model profile set as default — Opus during planning sessions, Sonnet during execution, Haiku for quick lookups

Lessons

  • The skill system makes Claude Code consistent across sessions — without it, every session starts from scratch
  • CLAUDE.md specificity is the multiplier: vague global instructions produce vague behavior; specific context produces specific output
  • Model selection is a workflow decision, not just a cost tradeoff — Opus for architecture, Sonnet for building, Haiku for trivia
  • /save as a closing habit is what makes everything else accumulate value over time