agentleFS
Sign inStart free
Dragoon0xDragoon0x/dragoon-skills/SKILL.md

useforgekit

Skill4 starsChanged 6 months agoMITView on GitHub

Generate optimal Claude Code configuration for any project. Creates CLAUDE.md, settings.json, skills, hooks, agents, and commands tailored to your stack. Run /forge to start.

  • Reads credentials

We split this skill into the rules it gives an agent

7 rules. About 0 are common, 4 are distinctive, and 3 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈4

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 3

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Run /forge when starting a new project or improving an existing Claude Code setup.

    line 11
  2. Very specific:

    Check for an existing .claude/.forge.lock — if present, switch to rerun mode

    line 32
  3. Distinctive:

    Create skill stubs for your domain

    line 35
  4. Distinctive:

    Run a health check on the complete setup

    line 38
  5. Distinctive:

    Read the project root. Look for:

    line 44
  6. Distinctive:

    If none found, ask: "What are you building?"

    line 62
  7. Very specific:

    After generating, run the health check (references/HEALTH-CHECK.md) to verify everything is correctly configured.

    line 107

≈ “Common” is an estimate: at least 1 in 5 of the 308,232 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

useForgeKit

When to Use

Run /forge when starting a new project or improving an existing Claude Code setup. This skill generates the complete configuration: CLAUDE.md, settings.json, skills structure, hooks, agents, and commands — all tailored to your specific stack and workflow.

Also use when:

  • Setting up Claude Code for the first time on a project
  • Onboarding a team to Claude Code
  • Auditing an existing setup for gaps
  • Migrating from another AI coding tool
  • Re-running on a project that already has forge config (see Rerun below)

Quick Start

/forge

Claude will:

  1. Detect your project type (or ask)
  2. Check for an existing .claude/.forge.lock — if present, switch to rerun mode
  3. Generate or update CLAUDE.md tailored to your stack
  4. Generate or update settings.json with appropriate permissions and hooks
  5. Create skill stubs for your domain
  6. Set up hooks for quality enforcement
  7. Configure agents if your project needs them
  8. Run a health check on the complete setup

The Setup Process

Step 1: Project Detection

Read the project root. Look for:

FileIndicates
package.jsonNode.js / JavaScript / TypeScript
tsconfig.jsonTypeScript specifically
next.config.*Next.js
vite.config.*Vite
requirements.txt / pyproject.tomlPython
Cargo.tomlRust
go.modGo
GemfileRuby
pom.xml / build.gradleJava / Kotlin
.swift filesSwift / iOS
docker-compose.ymlContainerized
supabase/Supabase backend
.github/workflows/GitHub Actions CI/CD

If none found, ask: "What are you building?"

Step 2: Generate Configuration

Load the appropriate reference module based on detection:

ReferenceWhen to Load
references/CLAUDE-MD.mdAlways — CLAUDE.md structure and content
references/SETTINGS.mdAlways — settings.json configuration
references/SKILLS-ARCHITECTURE.mdWhen creating skills
references/HOOKS.mdAlways — hook setup
references/AGENTS.mdWhen project needs subagents
references/COMMANDS.mdAlways — slash command setup
references/PERMISSIONS.mdAlways — permission strategy
references/CONTEXT-WINDOW.mdAlways — token budget management
references/PROJECT-TEMPLATES.mdBased on detected stack
references/ANTI-PATTERNS.mdAlways — what to avoid
references/DECISION-TREE.mdAlways — what goes where
references/WORKFLOWS.mdBased on project complexity
references/GIT-STRATEGY.mdAlways — what to commit
references/TEAM-SETUP.mdWhen team size > 1
references/PERFORMANCE.mdAlways — speed optimization
references/MCP-GUIDE.mdWhen external tools needed
references/TROUBLESHOOTING.mdWhen auditing existing setup
references/HEALTH-CHECK.mdAlways — final validation
references/RERUN-SEMANTICS.mdAlways — region markers, lockfile, diff flow
references/PROMPTING.mdPrompt patterns, challenge prompts, voice triggers, anti-patterns
references/SUBAGENT-PATTERNS.mdCommand→Agent→Skill, parallel agents, teams, worktrees, memory
references/DEBUGGING-WITH-CLAUDE.mdIron law, 3-strike rule, systematic methodology, scope freezing
references/TESTING-STRATEGY.mdTest priority, AAA pattern, edge cases, coverage targets, TDD
references/SECURITY.mdSecret management, code security checklist, supply chain, git security
references/RULES-GUIDE.md.claude/rules/ files, when rules vs skills vs CLAUDE.md
references/MEMORY.mdAuto-memory, managing memories, memory vs CLAUDE.md
references/PLUGINS.mdInstalling, creating, recommending plugins
references/ENVIRONMENT.md.env hierarchy, env var strategy, secrets
references/ADVANCED-HOOKS.mdChained hooks, conditional hooks, Stop/SessionStart/PreCompact hooks
references/MONOREPO.mdCLAUDE.md hierarchy, workspace commands, package-level skills
references/CI-CD.mdNon-interactive mode, CI workflows, cost control
references/COST-OPTIMIZATION.mdToken costs, model routing, budget estimation
references/ERROR-RECOVERY.mdRecovery playbook, safety stack, post-incident checklist
references/MIGRATION.mdFrom Cursor, Copilot, Windsurf to Claude Code

Step 3: Health Check

After generating, run the health check (references/HEALTH-CHECK.md) to verify everything is correctly configured.

Rerun

/forge is safe to run more than once. When .claude/.forge.lock exists, forge switches to rerun mode and respects the contract:

  • forge owns scaffolding — content inside <!-- forge:start id=... --> / <!-- forge:end --> markers in CLAUDE.md, and keys listed in owned_keys in settings.json.
  • you own decisions — anything outside those markers, and any key not in owned_keys. Forge never touches it.
  • the lockfile mediates — if forge-owned content has been edited (hash differs from lockfile), forge surfaces a diff instead of overwriting.

Modes:

  • /forge — default. Diff prompts on conflict.
  • /forge --dry-run — show every change, write nothing.
  • /forge --force — regenerate everything. Backs up to .forge-backup-{timestamp}/ first.
  • /forge --audit — read-only drift report. Useful for "is my setup still good?"

Full algorithm, lockfile schema, and migration rules: references/RERUN-SEMANTICS.md.

Output Structure

your-project/
├── CLAUDE.md                         # Project config (with forge regions + your notes)
├── .claude/
│   ├── .forge.lock                   # Lockfile — commit this
│   ├── settings.json                 # Hooks, permissions, env
│   ├── settings.local.json           # Personal overrides (gitignored)
│   ├── skills/
│   │   └── [project-specific]/
│   │       └── SKILL.md
│   ├── agents/
│   │   └── [role-specific].md
│   └── commands/
│       └── [workflow].md
└── .gitignore                        # Updated with Claude-specific entries

useForgeKit by Dragoon0x. Educational and experimental. DYOR.

Download now (ZIP)

  • Other versions1 project uses a version of this.

    Word for word · 1

    1. Dragoon0x/useforgekit★ 1100% same

    Edited · 0

    None found.

    Wrote their own version · 0

    None found.

    How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free