agentleFS
grezxunegrezxune/ai-tools/skills/standards/SKILL.md

standards

Skill0 starsChanged 18 days agoMITView on GitHub

Default stack, code, UI, and security rules. Use before writing app code in a repo with no AGENTS.md.

11 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use spec and design.

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.
  • SentrySkill

    spec

    Produce a verified spec — problem, desired outcome, and acceptance criteria — before building. Use when given a GitHub/Linear issue link, when starting from a fuzzy in-conversation idea, or asked to "figure out what to do" / "scope this" / "what should I build" — whenever the *what* isn't yet pinned down. Grills the ambiguities, checks the premise against the code, and produces acceptance criteria you sign off on — then hands to design-first.

    Also in 3 projects that use this

  • Microsoft AzureSkill

    design

    Design rules and conventions for the API Center Portal. Use when editing detail pages, card components, badges, layout, navigation, sidebars, or headers. Covers badge styling, metadata placement, page navigation patterns, and layout alignment.

    Also in 3 projects that use this

The file

We split this skill into the rules it gives an agent

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

  • 0

    Common

    Most of the 8 edited versions we read keep these.

  • 4

    Distinctive

    This author's own choices. Worth a look.

  • 5

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  • could be automatic: a setting could enforce it
  1. Very specific:

    Never use @ts-nocheck. Fix the types.

    line 20
  2. Distinctive:

    After every change, run build, typecheck, and lint for what you touched, plus targeted tests for the changed behavior.

    line 23
  3. Distinctive:

    Every app is built from its own component library, so a change to one component reaches every screen.

    line 27
  4. Distinctive:

    Never use a scroll indicator, animated dots, or "scroll" hint text.

    line 36
  5. Very specific:

    Never allow horizontal page scroll.

    line 37
  6. Very specific:

    Never use Tailwind's ! important modifier.

    line 38
  7. Very specific:

    Never use alert(), confirm(), or prompt().

    line 39
  8. Distinctive:

    No hardcoded secrets, keys, or credentials.

    line 43
  9. Very specific:

    Never trust a client-supplied userId, email, or role for authorization.

    line 45

“Common” is measured: kept by at least 6 in 10 of the 8 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

Standards

My defaults for web apps. A repo's own AGENTS.md replaces them.

Stack

  • Package manager is bun.
  • Next.js App Router only. Never create a pages/ directory.
  • Server components by default. Client components only where interactivity requires it.
  • Give each entity state its own route: create/, [id]/, [id]/update/, [id]/manage/. Do not multiplex those states inside one route.
  • Convex for data and file storage. Better Auth for auth, Google as the initial provider, joined to Convex through the @convex-dev/better-auth component. It replaced Auth.js, which gets only security fixes since September 2025. Stripe for payments, server-side only, Connect for multi-party payouts. Resend for email.

Code

  • Never use @ts-nocheck. Fix the types.
  • Target 200 lines per file. Exceed it when splitting would hurt cohesion, and say why.
  • Import everything explicitly.
  • After every change, run build, typecheck, and lint for what you touched, plus targeted tests for the changed behavior. Never merge with failing tests.

UI

Every app is built from its own component library, so a change to one component reaches every screen.

  • Primitives live in src/components/ui/ and own every class and token: layout (stack, row, grid, split), text and headings, controls, feedback, tables, and the app shell.
  • Screens and domain components compose primitives. Outside src/components/ui/, lint fails className, style, and CSS imports with react/forbid-dom-props, react/forbid-component-props, and no-restricted-imports.
  • Primitive props take named options such as tone="muted", columns="fields", or gap={4}, never raw classes or values.
  • A missing piece becomes a new primitive or a new variant of one. A pattern that shows up twice becomes a primitive, and both uses move onto it.
  • /dev/ui demos every shared component in each of its states, the way screens use it, and is a 404 outside development. A test fails when a component has no demo. A component and its demo change in the same commit.
  • Tokens for color, type, spacing, radius, shadow, and motion live in globals.css. The repo's DESIGN.md, or UxStyle.md in an older repo, sets the direction and names the tokens.
  • Support dark and light themes.
  • Never use a scroll indicator, animated dots, or "scroll" hint text.
  • Never allow horizontal page scroll. Constrain layouts and set overflow-x: hidden at the root.
  • Never use Tailwind's ! important modifier. Fix the layout or hierarchy instead.
  • Never use alert(), confirm(), or prompt(). Use the project's branded modal.

Security

  • No hardcoded secrets, keys, or credentials. Environment variables only.
  • Validate tokens server-side. Never trust client state.
  • Never trust a client-supplied userId, email, or role for authorization. Derive identity from the verified server session, then enforce ownership against stored records.
  • Deny by default. Missing auth, missing ownership, or identity mismatch fails immediately.
  • Filter on the server. Pass each view's filters, search, sort, and page as query arguments, and return only the rows and fields the current user needs for that view. A client never receives a full collection to filter, sort, or trim itself.
  • Validate and sanitize untrusted input at every boundary: API, server actions, database writes, webhooks.
  • Log errors with enough context to debug and no sensitive payloads.

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.