/ship

description: Run all auditors and produce a ship-readiness scorecard

/ship — Pre-Deploy Quality Gate

Run all Ultraship auditors in parallel and produce a screenshot-shareable scorecard.

What It Runs

5 scored categories from 6 tools, all in parallel:

| Category | Tools | What it checks | |---|---|---| | **SEO + AI Visibility** | seo-scanner | 39 SEO rules (meta tags, headings, OG tags, canonicals, cross-page analysis) + 20 GEO rules (AI bot access, snippet restrictions) + 4 AEO rules (schema checks) | | **Accessibility** | a11y-scanner | WCAG 2.2 A/AA static checks: missing alt text, unlabeled form controls, icon-only buttons, missing `lang`, heading order, positive tabindex, zoom disabled, broken aria references | | **Security** | secret-scanner | AWS keys, Stripe keys, GitHub tokens, private keys, DB URLs, JWT secrets in source files | | **Code Quality** | code-profiler + dep-doctor | N+1 queries, sync I/O in handlers, memory leaks, unbounded queries, unused/outdated dependencies | | **Bundle Size** | bundle-tracker | Build output size, heavy dependency detection (moment→dayjs, lodash→native, axios→fetch) |

How to Run

**In Claude Code:**

/ship

**Standalone CLI:**

npx ultraship ship .
npx ultraship ship /path/to/project

Scoring

• Each category starts at 100, deducts per finding based on severity
• SEO deduplicates by rule (same issue on N pages = one deduction)
• Failed tools show as `FAIL` and are excluded from the overall average
• Overall = average of categories that successfully ran

**Status thresholds:**

• >= 80: READY TO SHIP
• 60-79: NEEDS WORK
• < 60: NOT READY

After the Scorecard

The scorecard tells you what's wrong. To fix issues:

1. Run individual audit commands for detailed findings:

• `ultraship seo .` — full SEO findings with file:line locations
• `ultraship a11y .` — all WCAG 2.2 accessibility findings
• `ultraship security .` — all detected secrets
• `ultraship profile .` — N+1 queries, sync I/O, memory leaks
• `ultraship deps .` — unused and outdated packages

2. In Claude Code, use the corresponding skills which include fix guidance:

• `/seo` — SEO audit with fix suggestions
• `/a11y` — accessibility audit + auto-fix
• `/secure` — security audit with fix suggestions
• `/profile` — code quality fixes

3. Re-run `/ship` to verify your score improved.

Key Principles

• **Detect and report honestly** — tools find issues, Claude helps fix them
• **Never block on missing tools** — if a tool fails, score excludes it and shows FAIL
• **Scorecard is evidence** — screenshot-shareable proof of ship-readiness