One CLI to scaffold, audit and fix your repo's whole toolchain — linting, tests, commits, releases & CI.

Most tooling libraries give you one piece — just TypeScript configs, or just an ESLint preset. repo-tooling covers the entire lifecycle: TypeScript, Biome/ESLint, Vitest/Jest, Commitlint, Husky, Semantic Release, GitHub Actions CI, and supply-chain security (Dependabot + CodeQL) — all wired together. The interactive setup wizard scaffolds everything in one shot; doctor checks an existing project for drift; fix applies the missing pieces incrementally.

Full documentation →

Package manager: pnpm, by design. Every generator scaffolds pnpm workflows, workspace files, and scripts. repo-tooling targets pnpm only for now — npm, yarn, and Bun aren't generated.

Start a new project

Interactive wizard — answers every prompt, scaffolds the whole project:

npx @rtorcato/repo-tooling setup

Non-interactive — scaffold from a named preset in one shot (CI-friendly):

npx @rtorcato/repo-tooling setup --preset library -d ./my-lib --skip-install
# presets: library | web-app | node-api | nextjs-app | react-app | swift-library

swift-library scaffolds a SwiftPM package (manifest, sources, tests, SwiftLint, Periphery, macOS CI) instead of an npm one — see the Swift guide.

Just one config file? Use copy:

npx @rtorcato/repo-tooling copy biome        # → biome.json
npx @rtorcato/repo-tooling copy tsconfig     # → tsconfig.json
npx @rtorcato/repo-tooling copy changesets   # → .changeset/config.json
npx @rtorcato/repo-tooling copy oxlint       # → .oxlintrc.json
npx @rtorcato/repo-tooling copy claude-skill # → .claude/skills/repo-tooling.md

The Biome preset honours your .gitignore, so build output stays unlinted. One known gap: Biome cannot parse Tailwind CSS v4 stylesheets (@theme, @source, @custom-variant, @apply) and reports a parse error on them. If you use Tailwind v4, exclude your stylesheet in biome.json:

{
  "extends": ["@rtorcato/repo-tooling/biome"],
  // `includes` replaces the preset's list rather than extending it — restate
  // the exclusions you still want alongside the CSS one.
  "files": { "includes": ["**", "!**/node_modules", "!**/dist", "!**/*.css"] }
}

Already have a project? Don't rerun setup — use doctor + fix:

npx @rtorcato/repo-tooling doctor   # find what's missing or drifted
npx @rtorcato/repo-tooling fix      # apply scaffolders, prompting per item

See the Getting Started guide for the full walkthrough.

Commands

| Command | What it does | Example | | --- | --- | --- | | setup | Interactive wizard that scaffolds a whole new project (add --preset to run non-interactively). | npx @rtorcato/repo-tooling setup | | list | List every tooling configuration this package can scaffold (--json for machine output). | npx @rtorcato/repo-tooling list | | copy <config> | Copy a single config file into the current project. | npx @rtorcato/repo-tooling copy biome | | doctor | Diagnose an existing project for missing or drifted tooling. | npx @rtorcato/repo-tooling doctor | | fix [target] | Apply scaffolders for what doctor flagged (--yes, --dry-run, --diff). | npx @rtorcato/repo-tooling fix |

Prefer to run the audit in CI? doctor also ships as a GitHub Action:

- uses: rtorcato/[email protected]

See the GitHub Actions reference for its inputs and outputs.

Every command takes -d, --directory <path>; run any with --help for its full flags. Run list (or list --json) for the full set of fix targets — it's the source of truth. Notable ones include fix docs-site (scaffold a Docusaurus docs site) and fix bun (Bun runtime config).

The .repo-tooling.json lockfile

An optional manifest that records the tooling choices you adopted. Nothing reads it at build/lint/test time — it exists only so doctor can tell an intentional opt-out from drift. repo-tooling works fine without it, which is why doctor reports a missing one as not configured, not an error.

Generate or refresh it from what's currently on disk:

npx @rtorcato/repo-tooling fix lockfile

It's the only config file you need. Don't keep a separate hand-authored .repo-tooling.config.json next to it — the lockfile already embeds the full ProjectConfig under config, and setup --config accepts the lockfile directly, so you can re-run a non-interactive setup from it:

npx @rtorcato/repo-tooling setup --config .repo-tooling.json

Each config.* key mirrors a setup answer — see the schema for the full field reference. The keys doctor acts on:

• typescript.enabled / typescript.config — base | react | next | node | express
• linting.tool — biome | eslint | both | none
• formatting.tool — biome | prettier | none
• testing.framework — vitest | jest | playwright | none
• gitHooks, commitLint, semanticRelease — booleans
• securityAutomation — boolean (Dependabot + CodeQL)
• bundler — tsup | esbuild | vite | none
• aiSetup — boolean (AGENTS.md, CLAUDE.md, Cursor/Copilot rules)

How opt-outs actually work. When the lockfile records that you declined an optional tool (e.g. securityAutomation: false), doctor demotes that check from not configured to ok — intentionally declined instead of nagging you to add it. This applies only to optional checks — it does not silence drift: if a config you adopted diverges from the shared base (say your tsconfig.json extends a different preset), doctor still flags it, because drift is a mismatch in a tool you're using, not an opt-out. There's no field to suppress drift today.

Adoption vs. standalone. fix lockfile infers adoption from what's on disk, so on a repo that deliberately runs standalone configs (no @rtorcato/repo-tooling dependency) it can record typescript.config: "base" and friends — asserting you extend the shared bases when you don't. If you're not adopting repo-tooling's configs, skip the lockfile; it won't stop the extends drift warnings.

AI agent rules

The package ships rules that teach AI coding agents to drive the CLI (doctor / fix / setup) non-interactively. Install for your agent — all generated from one source, so guidance never drifts between them:

npx @rtorcato/repo-tooling fix claude-skill --yes           # → .claude/skills/repo-tooling.md
npx @rtorcato/repo-tooling fix cursor-rules --yes           # → .cursor/rules/repo-tooling.mdc
npx @rtorcato/repo-tooling fix copilot-instructions --yes   # → .github/copilot-instructions.md
npx @rtorcato/repo-tooling fix agents-md --yes              # → AGENTS.md

copilot-instructions and agents-md upsert a delimited block, so your own content in those shared files is never clobbered. Re-running updates the block in place on upgrade.

Prefer a symlink that auto-syncs the Claude skill on every upgrade?

mkdir -p .claude/skills
ln -sf ../../node_modules/@rtorcato/repo-tooling/tooling/claude/repo-tooling.md \
  .claude/skills/repo-tooling.md

Use with Claude Code (plugin)

This repo is also a self-hosted Claude Code marketplace. Install the plugin to get three skills — repo-tooling (adopt/audit the presets via the CLI), npm-publish (the family's release rules) and ai-issue-loop (the label-driven issue → PR pipeline) — in any session:

/plugin marketplace add rtorcato/repo-tooling
/plugin install repo-tooling@repo-tooling

ai-issue-loop also installs on its own, user-globally, so every repo on the machine shares one copy:

npx @rtorcato/repo-tooling fix claude-skills   # → ~/.claude/skills/ai-issue-loop/SKILL.md

It writes outside the repo, so it is opt-in: a bare fix skips it. See the AI Issue Loop guide.

Use with other AI tools (Cursor / Copilot / Codex)

AGENTS.md at the repo root carries the same guidance in the cross-tool convention many agents read, and ships in the npm tarball so tools scanning node_modules/@rtorcato/repo-tooling can find it.

What's new

See CHANGELOG.md for the full history.

Related packages

• @rtorcato/js-common — General TypeScript/JS utilities (strings, dates, numbers, async, errors)
• @rtorcato/browser-common — Browser Web API wrappers (clipboard, observers, storage, etc.)

Roadmap

Direction and progress are tracked entirely on GitHub — see the milestones and open issues.

Contributing

Contributions welcome — see CONTRIBUTING.md.

License

MIT — see LICENSE.

Install the skills (npx skills)

Any agent that supports the skills CLI can install this repo's skills straight from GitHub — no clone, no package install:

npx skills add https://github.com/rtorcato/repo-tooling --skill ai-issue-loop
npx skills add https://github.com/rtorcato/repo-tooling --skill npm-publish
npx skills add https://github.com/rtorcato/repo-tooling --skill repo-tooling