agentleFS
denniemokdenniemok/simple-skills/simple-launch/SKILL.md

simple-launch

Skill1 starsChanged 43 days agoMITView on GitHub

Pre-launch checklist for a website or web app, covering metadata, SEO, social cards, content, performance, and production config. Use only when the user asks to prepare for launch.

  • Reads credentials

The file

We split this skill into the rules it gives an agent

8 rules. About 0 are common, 7 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈7

    Distinctive · estimate

    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:

    Ask for the production URL if the config does not have the real one.

    line 18
  2. Distinctive:

    If you find one:

    line 37
  3. Distinctive:

    Never print the full value.

    line 38
  4. Distinctive:

    After rotating, update every deployment that used it.

    line 40
  5. Very specific:

    No placeholder text, lorem ipsum, TODO, or test data

    line 71
  6. Distinctive:

    Run Lighthouse on the built site, mobile first.

    line 79
  7. Distinctive:

    Check the main pages at phone, tablet, and desktop widths

    line 82
  8. Distinctive:

    Test the social card with the sharing debugger of each platform

    line 95

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 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

Simple Launch

Check a site is ready to go live. Read the repo and the built output, then report each item as pass, fail, or not applicable, with the file. Fix only when the user asks. Start with the secrets sweep, since a leak blocks everything else. For DNS, SSL, caching, and security headers, use simple-cloudflare. For the database, use simple-supabase.

Workflow

  1. Detect the framework (Astro, Next.js, Vite SPA, or other) and the hosting target.
  2. Ask for the production URL if the config does not have the real one.
  3. Go through the sections below.
  4. Report failures first, with the file and the fix. Then list what passed.

1. Secrets sweep

A leaked key is the highest impact failure, so treat any hit as a launch blocker. Search four places:

  1. Working tree. Look for known key shapes: sk-, sk_live_, AKIA, ghp_, github_pat_, xoxb-, AIza, -----BEGIN PRIVATE KEY-----, service_role, secret, password=, and connection strings with credentials such as postgres://user:pass@. Also check for JWTs starting with eyJ.
  2. Tracked files. Run git ls-files and look for .env, *.pem, *.key, credentials files, and database dumps. .gitignore must cover them, and .env.example must hold placeholders only.
  3. Git history. A secret deleted in a later commit is still in history. Use gitleaks detect or trufflehog git file://. if installed, otherwise git log -p -S "<pattern>".
  4. Built output. Search dist/, build/, or .next/static for the same patterns. Client-prefixed variables (VITE_, NEXT_PUBLIC_, PUBLIC_) are shipped to the browser, so none of them may hold a secret key. Check that production source maps are off, or do not expose anything sensitive.

Also check for:

  • Hardcoded admin or test accounts, debug flags, and open debug endpoints
  • Wildcard CORS (*) on an API that uses cookies or private data
  • Third-party keys meant for the browser (maps, analytics) restricted to your domain in that provider's settings
  • Secrets in the README, docs, screenshots, CI config, and Docker files

If you find one:

  • Never print the full value. Show the file, line, and type, with the value masked.
  • Tell the user to rotate the key first, then remove it. Removing it from the code alone does not make it safe.
  • After rotating, update every deployment that used it.

2. Identity and metadata

  • Each page has a unique <title> and meta description
  • lang is set on <html>
  • Canonical URL is set and uses the production domain
  • Favicon exists, with an Apple touch icon if the site is used on mobile
  • Social card: og:title, og:description, og:image (1200 x 630), twitter:card
  • The social image URL is absolute and loads

3. Discoverability

  • sitemap.xml is generated and lists the real URLs
  • robots.txt exists and does not block the whole site by mistake
  • Staging or preview builds are noindex, production is not
  • A custom 404 page exists, and old URLs have redirects
  • Semantic structure: one h1 per page, headings in order

4. Framework config

FrameworkCheck
Astrosite in astro.config is the production URL, since the sitemap and canonical URLs depend on it
Next.jsmetadataBase is set, app/sitemap.ts and app/robots.ts exist, metadata on each route
Vite SPAMeta tags in index.html, SPA fallback route on the host, and consider prerendering for SEO pages

Watch for leftover TODO, localhost, and default .github.io or .vercel.app URLs in the config.

5. Content

  • No placeholder text, lorem ipsum, TODO, or test data
  • Links and buttons work, and external links open safely
  • Images have alt text, correct sizes, and modern formats
  • Contact details, social links, and legal pages (privacy, terms) are correct
  • Copyright year and site name are right

6. Performance and accessibility

  • Run Lighthouse on the built site, mobile first. Aim for 90 or more in each category.
  • Images are sized and lazy loaded below the fold, fonts are preloaded or self-hosted
  • Keyboard navigation works, contrast is fine, forms have labels
  • Check the main pages at phone, tablet, and desktop widths

7. Production config

  • Production env vars are set on the host, and none are missing
  • HTTPS works and redirects from HTTP
  • Analytics is set up, plus a consent notice if the audience needs one
  • Error tracking or logs are reachable
  • The build succeeds from a clean install

8. After launch

  • Open the live site and click through the main flows
  • Test the social card with the sharing debugger of each platform
  • Submit the sitemap in Google Search Console
  • Set up uptime monitoring if the site matters

Output

  • Failures, each with file and fix
  • Warnings worth a look
  • Passed items in one line
  • Verdict: ready to launch, or not, and the blockers

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.