A Claude skill that audits your web app or SaaS product before launch. Give it your repo, your live URL, or both, and it returns a scored readiness report with every problem ranked by severity, exact fix instructions for each one, and copy-paste Claude prompts so you can resolve issues immediately without writing a single prompt yourself.
I built this because pre-launch checklists are either too generic to act on or scattered across a dozen blog posts. This skill pulls everything into one structured audit that runs inside Claude, reads your actual code or live URL, and tells you precisely what is broken and how to fix it.
I update this skill as new patterns, security issues, and launch gotchas emerge. All updates are published at omrajguru.com.
There are four ways to install depending on how you use Claude.
npx skills add omrajguru05/launch-readinessTo preview what will be installed before adding:
npx skills add omrajguru05/launch-readiness --listRun this in your terminal:
curl -fsSL https://raw.githubusercontent.com/omrajguru05/launch-readiness/main/install.sh | shThis downloads the skill directly into ~/.claude/skills/app-launch-readiness/. Restart Claude Code and the skill is live.
To update to the latest version at any time, run the same command again. The installer replaces the old version automatically.
Clone the repository and copy the skill folder into your Claude skills directory:
git clone https://github.com/omrajguru05/launch-readiness.git
cp -r launch-readiness/skills/app-launch-readiness ~/.claude/skills/app-launch-readinessTo install it scoped to a single project instead of globally:
mkdir -p .claude/skills
cp -r launch-readiness/skills/app-launch-readiness .claude/skills/app-launch-readinessRestart Claude Code after copying.
- Download
app-launch-readiness.skillfrom this repository. - Open claude.ai and go to Settings.
- Navigate to Customize, then Skills.
- Upload the
.skillfile. - Toggle the skill on.
Code execution must be enabled under Settings, then Capabilities for skills to work.
Once installed, trigger the skill by describing what you want audited. Any of these work:
- "Check if my app is ready to launch" followed by your GitHub URL or live URL
- "Run a pre-launch audit on this repo: github.com/yourname/yourapp"
- "Audit my app at https://yourapp.com before I ship"
- "What is missing before I can launch? Here is my repo..."
- "Is my SaaS ready to go live?"
The skill accepts three types of input:
A GitHub URL. Claude reads your key files including package.json, auth files, payment files, routing files, config files, and any policy pages it can find.
A live URL. Claude fetches the page, inspects response headers, checks for security headers, reads meta tags, looks for robots.txt and sitemap.xml, and verifies policy pages are reachable.
Both together. This gives the most complete audit. The codebase surfaces things invisible from outside such as secrets handling, webhook logic, and environment separation, while the live URL covers things that only exist at runtime such as response headers, rendered HTML, and cookie banners.
If you have neither ready, describe your stack, auth setup, payment provider, and hosting. The skill audits what it can and clearly flags what requires access to verify.
The skill evaluates your app across nine categories, each weighted by how much it affects launch viability.
| Category | Weight | What It Covers |
|---|---|---|
| Security | 20% | HTTPS, security headers, MFA, RBAC, secrets management, input validation, rate limiting, dependency CVEs |
| Legal and Compliance | 15% | Privacy Policy, Terms of Service, Cookie Consent, GDPR/CCPA, Refund Policy, AUP |
| Infrastructure | 15% | CDN, uptime monitoring, error tracking, database backups, CI/CD, environment separation |
| Payments and Billing | 15% | Stripe/LemonSqueezy integration, webhook handling, tax, dunning, proration, cancellation |
| SEO | 10% | Meta tags, Open Graph, sitemap, robots.txt, Core Web Vitals, schema markup, mobile responsiveness |
| Visual Polish | 10% | Screenshots, OG image, demo video, favicon, landing page structure, dark mode |
| Analytics and Monitoring | 5% | Event tracking, server logs, alerting on errors and traffic spikes |
| Social and Marketing | 5% | Handle claims, pre-launch content, community identification |
| Post-Launch Readiness | 5% | Support channel, feedback mechanism, rollback plan, on-call coverage |
Payments is marked N/A for free apps and its weight is redistributed proportionally across the other categories.
Every audit produces the same structured output.
Overall score on a 0 to 100 scale:
| Score | Status |
|---|---|
| 90 to 100 | LAUNCH READY |
| 75 to 89 | NEARLY READY |
| 50 to 74 | NEEDS WORK |
| Below 50 | NOT READY |
Per-category scores in a table so you can see exactly which area is pulling your overall score down.
Critical blockers listed first. These are FAIL items from Security and Legal. Each entry includes the specific problem observed in your code or URL, the impact of shipping without fixing it, exact steps to resolve it, and a ready-to-use Claude prompt you can paste into a new session to get working code immediately.
High priority issues from Infrastructure, Payments, and SEO in the same format.
Medium priority issues from the remaining categories in the same format.
What looks good, grouped by category, so you know what is already solid.
After the report, the skill offers to auto-fix the top fixable issues right in the same session, writing production-quality code or config changes for your specific stack.
Overall = (
Security * 0.20 +
Legal * 0.15 +
Infrastructure * 0.15 +
Payments * 0.15 +
SEO * 0.10 +
Visual * 0.10 +
Analytics * 0.05 +
Social * 0.05 +
PostLaunch * 0.05
)
Each item in a category is scored PASS, FAIL, WARN, or N/A. Category score equals PASS / (PASS + FAIL + WARN) * 100. N/A items are excluded from the denominator.
Every problem in the report includes a copy-paste Claude prompt. These prompts contain the framework and version, the current state of the missing item, exactly what needs fixing, and any constraints such as third-party services or existing config.
Example fix prompt the skill generates:
I am building a Next.js 14 app deployed on Vercel. My next.config.js currently
has no security headers configured. Please add a complete headers() export to
my next.config.js that includes Content-Security-Policy, X-Frame-Options,
Strict-Transport-Security, X-Content-Type-Options, Referrer-Policy, and
Permissions-Policy. The app uses Stripe for payments and loads fonts from
Google Fonts. Make the CSP permissive enough not to break those but secure
otherwise.
Paste this into a new Claude session and get working code with zero additional context required.
- Claude Pro, Max, Team, or Enterprise plan
- Code execution enabled under Settings, then Capabilities
I maintain this skill as new security patterns, infrastructure defaults, and launch requirements emerge. All updates are published at omrajguru.com.
To update to the latest version, run the install command again:
curl -fsSL https://raw.githubusercontent.com/omrajguru05/launch-readiness/main/install.sh | shWatch this repository to get notified when new versions ship.
launch-readiness/
├── README.md
├── LICENSE
├── install.sh
├── app-launch-readiness.skill
└── skills/
└── app-launch-readiness/
└── SKILL.md
A Claude skill is a .skill file or folder that packages a set of instructions, scoring logic, output formats, and edge case handling into a single installable unit. Skills are more reliable than one-off prompts because the entire workflow is defined once and applied consistently every time the skill triggers. Claude loads skills automatically when your request matches the skill description.
MIT. Use it, fork it, modify it. If you improve it significantly, a mention or a link back to omrajguru.com is appreciated.