Disclaimer: Educational and experimental. DYOR. See DISCLAIMER.md.
The definitive Claude Code setup skill. Install it once. Run /forge. Get a
complete, tailored configuration for your specific project: CLAUDE.md, settings.json,
skills, hooks, agents, and commands — all generated from your actual codebase.
Every other setup resource is a guide you READ. This is a skill Claude EXECUTES.
git clone https://github.com/Dragoon0x/useforgekit.git ~/.claude/skills/useforgekit/forge
Claude detects your stack, asks a few questions if needed, and generates everything.
- CLAUDE.md — Project context, conventions, gotchas (< 200 lines, tailored to your stack)
- settings.json — Permissions, hooks, model selection, thinking mode
- Skills — Domain-specific knowledge for your frameworks and patterns
- Hooks — Auto-format, branch protection, destructive command blocking
- Agents — Code reviewer (read-only), test generator, doc writer
- Commands — /fix-issue, /create-pr, /review, /deploy, /onboard
| Module | What It Covers |
|---|---|
| CLAUDE-MD.md | How to write CLAUDE.md: structure, sizing, anti-patterns, per-stack templates |
| SETTINGS.md | Every settings.json field: permissions, hooks, model, tokens, effort |
| SKILLS-ARCHITECTURE.md | Skill anatomy, frontmatter, progressive disclosure, categories |
| HOOKS.md | All hook events, matchers, output protocol, per-language auto-formatters |
| AGENTS.md | Agent anatomy, frontmatter, essential agents, Agent() invocation |
| COMMANDS.md | Command anatomy, $ARGUMENTS, essential commands |
| DECISION-TREE.md | What goes in CLAUDE.md vs skill vs agent vs command vs hook vs rule |
| CONTEXT-WINDOW.md | Token costs, budget management, compaction strategy |
| ANTI-PATTERNS.md | 30 enumerated anti-patterns with symptoms and fixes |
| PROJECT-TEMPLATES.md | Stack-specific configs: TypeScript, Python, Next.js, Rust, Go |
| WORKFLOWS.md | Standard dev loop, feature dev, bug fix, review, refactoring, debugging |
| GIT-STRATEGY.md | What to commit, .gitignore additions, branch strategy |
| PERFORMANCE.md | Speed tips, token efficiency, model selection by task |
| PERMISSIONS.md | Permission spectrum, recommended strategies, per-agent permissions |
| MCP-GUIDE.md | Essential MCP servers, configuration, MCP vs CLI, MCP in agents |
| TEAM-SETUP.md | Multi-developer config, onboarding, conflict resolution |
| TROUBLESHOOTING.md | Common issues with causes and fixes, /doctor |
| HEALTH-CHECK.md | 30-item validation checklist with verification commands |
MIT — see LICENSE.
By Dragoon0x. Zero dependencies. Educational and experimental.