agentleFS
gabimonchagabimoncha/skills/skills/essentials/project-setup/SKILL.md

project-setup

Skill12 starsChanged 20 days agoMITView on GitHub

Load this skill before you add a layer that a project lacks, such as auth or hosting, migrate a project to the user's stack, replace a layer with the stack's choice, or start a new project. Covers the stack table, what each choice replaces, the migration order, the new-project files, .worktreeinclude, the global package store, the file templates, and the done checks.

  • Reads credentials

52 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use creative-writing-craft, creative-writing-muse and 10 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • Abdullokh-NSkill

    creative-writing-craft

    Справочник ремесла: проза, сцена, стиль, голос. Загружать, когда нужно знание "как писать", а не режим производства.

    Also in 3 projects that use this

  • Abdullokh-NSkill

    creative-writing-muse

    Точка входа в набор скилов для длинной прозы. Включается, когда автор просит спланировать, написать, продолжить, переписать или разобрать художественный текст — роман, повесть, цикл рассказов. Ведёт маршрут от замысла к готовому куску и держит базу знаний проекта.

    Also in 3 projects that use this

  • badgidsSkill

    character-sim

    Pressure-test a character by speaking or reasoning strictly from that character's current voice, goals, relationships, emotional pressure, and knowledge boundary without creating canon by accident.

    Also in 3 projects that use this

  • jpsweeney97Skill

    writing-principles

    Enforces writing principles for instruction documents (SKILL.md, skill supporting files, subagents, CLAUDE.md). Use when writing, reviewing, or editing any file in **/skills/**/*.md, **/agents/*.md, or **/CLAUDE.md. Triggers on skill creation, skill file updates, CLAUDE.md changes, subagent authoring, or instruction document review.

    Also in 3 projects that use this

  • ItsSidharthaSkill

    story-planning

    Collaborative planning skill for a GitHub issue. Takes issue content and codebase context as input. Runs an iterative clarification loop until the user explicitly agrees to a plan, then saves it to docs/<story_number>/plan.md. Used by the user-story-implementation skill — load this skill at the start of Phase 1 and follow it completely before returning to the orchestrator.

    Also in 3 projects that use this

  • badgidsSkill

    story-review

    Diagnose existing prose at a chosen editorial level, from whole-story and developmental problems through scene, line, copy, and proof passes, without rewriting until the diagnosis is accepted.

    Also in 3 projects that use this

  • +4 more from Abdullokh-N

The file

We split this skill into the rules it gives an agent

5 rules. 0 are common, 4 are distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 44 edited versions we read keep these.

  • 4

    Distinctive

    This author's own choices. Worth a look.

  • 1

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    When a task needs a layer, or a part of one such as auth, that the project lacks, use the table's choice.

    line 34
  2. Distinctive:

    Keep the behavior the same, and keep fixes and features out of it.

    line 46
  3. Distinctive:

    After each layer, run the app and the checks.

    line 47
  4. Distinctive:

    Ask about layers the table doesn't cover.

    line 53
  5. Very specific:

    Start each file from its template in this skill's assets/ folder.

    line 91

“Common” is measured: kept by at least 6 in 10 of the 44 edited versions of this file we could read. “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

Project setup

The stack table is the target for the user's own projects. Use it for a layer that a project lacks, for a migration, and for a new project. A change inside a tool the table already names is ordinary work. Run only the layers the user asked for.

The stack

LayerChoiceReplaces
Packages and monorepoBun, bun.lock, workspace catalogs, mise tasks. Turborepo only for its cache or for runs in the changed packagesnpm, yarn; Turborepo for plain task runs
Web appTanStack Start, Vite, React, Tailwind 4, shadcn/uiNext.js
Mobile appExpo, Expo Router, dev client, EAS; Uniwind with DESIGN.md tokensNativeWind, StyleSheet
APIBun and Elysia for a long-running server; TanStack Start server routes for web app endpointsExpress, NestJS; Next.js API routes
Data and authConvex with Clerk; for SQL, Drizzle with Neon and Better AuthFirebase, Supabase
HostingCloudflare (Workers, Pages, R2) through Alchemy; Coolify for DockerVercel
Code checksoxlint, oxfmt, tsrs (TypeScript 7), hk hooks, fallow, react-doctorESLint, Prettier, Biome; tsc, tsgo; lefthook, husky; knip
Formal verificationtla-ts, for TLA+ models of state machines and async flows
Testsbun test in Bun projects; otherwise Vitest. Vitest where convex-test needs it. e2e with agent-device (iOS) and agent-browser (web)Maestro, Playwright, Detox
CI and commitsGitHub Actions on Blacksmith runners, with mise-action; Conventional Commits; Renovate for dependency updatesGitHub-hosted runners
Env and secretsInfisical for environment variables; Varlock for agent access
Agent workspaceT3 Code, T3 worktrees, a t3.json per repo, and a .worktreeinclude

Add a missing layer

When a task needs a layer, or a part of one such as auth, that the project lacks, use the table's choice. Add only that layer, in the task's own PR.

Migrate a project

  1. Compare the project with the stack table. List what to replace, and what the project lacks, such as formal verification, the global package store, or .worktreeinclude. Give the order, then wait for the user's pick. Skip the wait when the user already named the layers.
  2. For a move into mise, load mise-review and stop at its plan. It is read-only. Do the move after the user approves the plan.
  3. Move one layer per PR. Use the versions that the global AGENTS.md sets. Keep the behavior the same, and keep fixes and features out of it.
  4. After each layer, run the app and the checks. Then remove the old tool completely: its config, scripts, and dependency. Remove the old lockfile only when the package manager changes.

Start a new project

  1. Ask about layers the table doesn't cover. Ask before you create the GitHub repo. The first commit on the default branch needs no question.
  2. From day one, the project has:
    • mise.toml with pinned tools, worktrunk among them, and setup and check tasks
    • repo config that turns on the global store of the package manager in use, so a new worktree or a teammate's clone doesn't download every package again
    • hk hooks and CI
    • .worktreeinclude: the gitignored files and folders that a new worktree copies from the main checkout. List the .env files, .claude/settings.local.json, mise.local.toml, plans/, .scratch/, and any other local file the work needs. Leave out a file that must differ per worktree, such as an env file that points at one backend deployment. Write a folder as plans/, never plans/*: Worktrunk copies whole ignored folders, so plans/* copies nothing, and a ! line inside a listed folder has no effect.
    • t3.json, with a Setup script that runs when T3 creates a worktree ("runOnWorktreeCreate": true). Its command is wt step copy-ignored --require-include && mise install && mise run setup. T3 doesn't read .worktreeinclude, so wt copies the listed files from the main checkout first. It never overwrites a file. Claude Code and the Codex app read .worktreeinclude themselves.
    • .github/pull_request_template.md, from the pr skill's template
    • AGENTS.md, and a CLAUDE.md that holds only @AGENTS.md
    • the other files in "Templates" that apply
  3. Make one real path work end to end. No sample pages, demo data, or placeholder code.
  4. It's done when mise run setup and mise run check pass. Check this on a fresh clone, and in a new worktree from delegate_task with isolation: "worktree". That worktree holds the .worktreeinclude files.

For a project with a runnable app, also add a dev task and one E2E check. It's done only when mise run dev serves the app and its ready check passes. A library, a template, or a skills repo has no dev task and no E2E check.

Templates

Start each file from its template in this skill's assets/ folder. Replace every <…> placeholder with a fact from the project. Delete a section or a line that has no fact yet. A file is done when no placeholder is left.

Project fileTemplateHolds
.worktreeincludeworktreeincludeThe local files that a new worktree copies from the main checkout.
t3.jsont3.jsonThe Setup script that T3 runs when it creates a worktree.
AGENTS.mdAGENTS.template.mdShort always-on rules, then pointers such as "Before you do X, read docs/agents/x.md".
docs/agents/*.mddocs/agents/How-tos for agents: git workflow, dev runtime, and verification. Add one only when the repo has facts for it.
GLOSSARY.mdGLOSSARY.mdDomain terms. Create it when it gets its first term. The domain-modeling skill owns its rules.
DESIGN.mdDESIGN.mdThe visual system of a project with UI.
PRODUCT.mdPRODUCT.mdProduct intent.

docs/adr/ holds decisions and has no template. The domain-modeling skill owns its format. Create the folder with the first decision. In a migration, GLOSSARY.md and docs/adr/ replace CONTEXT.md.

Download now (ZIP)

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.