Claude Code plugin with Currency Guard, elite SEO strategy (GSC + Bing + GA4 keyword intelligence, index doctor, AI traffic tracking, IndexNow, cross-reference, CTR anomalies, brand filtering), sprint workflow, safety guardrails, canary monitoring, retrospectives, pre-deploy auditing (SEO, security, penetration testing, code profiling, bundle size), workflow enforcement (TDD, code review, planning), and project tooling (competitive analysis, architecture mapping, codebase indexing, incident response). 41 tools, 45 skills, 13 agents.
Ultraship presents 26 tools with severe definition gaps. Most tool definitions are inferred from brief descriptions in package.json and bin/ultraship.mjs rather than explicit MCP tool registration. Input schemas are either missing or not visible in provided source. Descriptions exist but are often generic (e.g., 'Run SEO audit + AI visibility checks') without actionable guidance on WHEN to use each tool or WHAT parameters control behavior. No output schemas documented. Parameter descriptions are minimal or absent. Error handling and recovery guidance are not evident. The architecture-mapper tool shows a proper schema with 'project_directory' and 'format' params, but this is an exception, most tools lack visible schema definitions. Tools like 'incident-commander' and 'launch-prep' have vague descriptions that don't disambiguate them from overlapping tools (e.g., both could diagnose production issues). No evidence of permission gates, audit trails, or secret injection patterns.
Accessibility audit (WCAG 2.2). Scans for accessibility violations and compliance issues.
Living Architecture Map — auto-generates Mermaid diagrams from codebase analysis. Detects routes, database schemas, services, middleware, and import graphs. Identifies circular dependencies.
Analyze bundle size. Tracks and reports on JavaScript bundle size metrics and optimization opportunities.
Find performance anti-patterns. Analyzes code for performance bottlenecks, inefficient patterns, and optimization opportunities.
Generate compact codebase index (saves tokens). Creates a compressed, token-efficient representation of the codebase for LLM context.
Competitive X-Ray (2 URLs). Compares two competitor URLs across SEO, performance, features, and technology stack.
Score content quality. Evaluates content for readability, SEO optimization, and engagement metrics.
Missing or non-visible input schemas for 24 of 26 tools. Only architecture-mapper and index-doctor show explicit schema definitions. Most tools inferred from brief descriptions rather than explicit MCP tool registration.
Generic, non-actionable descriptions lacking WHEN/WHY guidance. E.g., 'Run SEO audit + AI visibility checks' does not explain when to call seo-scanner vs launch-prep vs demo-prep, or what parameters control behavior. Descriptions average ~50 chars; baseline is 194 chars.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 43 | 2026-07-28+ | v2 |
AI build cost tracking. Monitors and reports on AI/LLM usage costs during development.
Demo readiness check. Verifies that the application is ready for demonstration with proper data, performance, and UX.
Find unused/outdated dependencies. Analyzes package.json and identifies unused packages and outdated versions.
Validate environment variables. Checks for required env vars, type validation, and configuration completeness.
Find AI features (LLM call sites) that need an eval/regression harness. Identifies LLM integration points requiring evaluation.
Post-ship growth metrics. Tracks and reports on post-deployment growth metrics and KPIs.
Check production URL health. Verifies uptime, response times, and HTTP status codes.
Production incident diagnostics. Provides rapid diagnosis and remediation steps for production incidents.
Diagnoses and fixes non-indexed pages. Commands: diagnose (check indexing status of all sitemap URLs), inspect (deep inspection of single URL), fix (auto-submit non-indexed pages for re-indexing), coverage (GSC index coverage report), compare (compare GSC vs Bing indexing)
Launch day preparation. Comprehensive pre-launch checklist covering SEO, performance, security, and marketing readiness.
Check pending migrations. Identifies unapplied database migrations and schema drift.
Validate Open Graph tags. Checks for proper OG meta tags and social media preview optimization.
Generate onboarding guide. Creates comprehensive onboarding documentation for new team members.
Analyze codebase patterns. Identifies common patterns, anti-patterns, and architectural consistency.
Check redirect chains. Identifies redirect loops, excessive hops, and broken redirect destinations.
Scan for leaked secrets. Detects exposed API keys, credentials, and sensitive data in codebase.
Run SEO audit + AI visibility checks. Analyzes on-page SEO, structured data, Open Graph tags, Core Web Vitals, indexability, and AI traffic signals.
Blocking quality gate (exit 1 on fail). Enforces quality standards as a pre-deploy gate. Commands: init|run|ci|hook
Vibe-coding security: detects public-prefixed secrets, client service keys, and missing RLS (Row Level Security) policies.
No output schemas documented. LLMs cannot plan downstream tool calls or extract required fields. No evidence of pagination, result limits, or structured response formats.
Overlapping tool responsibilities without clear disambiguation. incident-commander, launch-prep, demo-prep, and health-check all appear to diagnose production readiness or issues. LLMs will struggle to select the right tool.
No evidence of error handling, recovery guidance, or permission gates. Tools like secret-scanner and vibe-security-scanner lack guidance on what to do if secrets are found. No audit trails or security patterns visible.