Skip to content

CLAUDE.md

Latest commit

 

History

History
51 lines (32 loc) · 3.59 KB

CLAUDE.md

File metadata and controls

51 lines (32 loc) · 3.59 KB

CLAUDE.md

This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.

What This Repo Is

This is Labrys' global Claude Code configuration. It is cloned to ~/.claude and provides coding standards, slash commands, and specialized agents that apply across all projects. It is not a software project — there is nothing to build, lint, or test.

Structure

  • rules/ — Coding standards auto-loaded by Claude Code into every session
  • skills/ — Slash commands (/pr, /check, /testing-plan) in the skills format
  • agents/ — Specialized subagents for domain-specific tasks
  • docs/ — Internal guides (Subagents Guide, Monorepos Guide)

Rules (auto-loaded)

These define the coding conventions enforced across all projects:

  • react.md — Controller-view-hook pattern: every feature splits into feature.tsx (thin controller), feature.hook.ts (all business logic), feature.view.tsx (pure presentation). Views may only hold ephemeral UI state (modals, dropdowns). Rule of thumb: "what" data → hook, "how" displayed → view.
  • typescript.md — any is forbidden; use unknown, generics, union types, or type guards instead. Document the rare exception.
  • style.md — PascalCase components/classes, camelCase functions, SCREAMING_SNAKE_CASE constants, hyphenated-lowercase file names (except component files which match their export). Simplicity over cleverness, explicit over implicit.
  • documentation.md — README must stay in sync with code; update immediately on architectural changes.

Skills (Slash Commands)

Skills live in skills/<name>/SKILL.md and are invoked as /name. All three are user-only (disable-model-invocation: true) — Claude will not auto-trigger them.

  • /pr [base-branch] — Analyzes diff, generates PR description, creates or updates PR via gh. Never pushes without permission.
  • /check — Discovers CI workflows in .github/workflows/, runs locally-runnable checks in parallel, auto-fixes format/lint issues, reports remaining problems.
  • /testing-plan <file-path>[#L<line>] — Generates an execution-ordered unit testing plan as a markdown file co-located with the source.

Agents

Two tiers of agents exist:

Generic agents (agents/): Reusable across any project — frontend-developer, backend-developer, fullstack-developer, nextjs-expert, typescript-pro, code-reviewer, debugger, documentation-engineer, qa-expert. Each encodes domain expertise and references the rules above.

Specialist agents (agents/example-specialists/): Project-specific examples showing how to create deep-domain agents. The included examples (lagoon-vault-analyst, lido-staking-vault-analyst) demonstrate the pattern — they encode protocol-specific contract architectures, flows, and analysis checklists. The generic defi-protocol-analyst agent routes to these specialists when appropriate.

Agent anatomy

Each agent is a markdown file with YAML frontmatter (name, description, model, color) followed by the agent's system prompt including expertise, patterns, tool usage guidance, and cross-references to related agents.

Editing Guidelines

When modifying this repo:

  • Rules should be concise and opinionated — state what to do, not every possible alternative
  • Skills should be self-contained instructions that work without user interaction where possible
  • Agents should encode domain knowledge that would take multiple files to discover, not generic advice
  • Keep specialist agents in agents/example-specialists/ to separate them from the generic set