When the user wants to audit, review, or diagnose SEO issues on their site. Also use when the user mentions "SEO audit," "technical SEO," "why am I not ranking," "SEO issues," "on-page SEO," "meta tags review," "SEO health check," "my traffic dropped," "lost rankings," "not showing up in Google," "site isn't ranking,"…
Add and maintain SEO-ready MDX articles — metadata, structured data, sitemap, OG images, internal linking. Use when writing a blog post or auditing a page's search surface.
Use when connecting, authorizing, provisioning, resuming, verifying, canceling, or diagnosing an external service such as Convex, Stripe, GitHub, Resend, PostHog, Netlify, Vercel, or Cloudflare, especially when browser consent or another human-only step must pause an agent running in Claude Code, Codex, Cursor…
This compatibility skill should be used when the user asks to "run setup health", "check every connection", "verify Agentic Ship setup", or invokes the former all-in-one health workflow; it routes the request to workspace-health, agent-compatibility, service-connections, upstream-sync, or production-preflight without…
The verification gates, test-data rules, and the self-healing loop — unit tests with vitest/convex-test, e2e with Playwright, tier-1 deterministic repair via pnpm heal, and the healer guardrails. Use when writing tests, when any gate is red, or when a repair is needed.
Establish and enforce the project's visual system — design tokens, typography, radius, spacing, dark mode — and route component sourcing and asset work. Use when starting a project, changing the theme, adding a new interface piece or visual asset, or when generated UI starts looking like every other AI-built site.
Check every vendored skill, MCP server, registry, and version pin against its official upstream; report drift; apply clean updates; three-way merge anything patched. Run monthly, or when a tool ships a major version.
Turn product intent and attributed visual references into a validated visual plan, then review implemented UI for generic generated-page residue. Use before creating a route, page, section, component family, theme, or substantial visual revision, and when an existing interface feels generic or AI-generated.
Capture, inspect, repair, and accept deterministic visual evidence after authored UI changes. Use before declaring UI complete, when screenshots or UI evidence are missing or stale, when responsive/theme/state behavior needs review, or when a visual gate is red.
This skill should be used when the user asks to "check workspace health", "verify the local setup", "diagnose generated configuration drift", "check MCP or agent mirrors", "audit component structure", or investigate a failing deterministic repository gate without authorizing or provisioning an external service.
Write and review documentation, README content, wiki articles, and any reader-facing prose in this repository or a product workspace. Use when asked to "write docs", "write an article", "build a wiki page", "review my docs", "check writing style", or when finishing any document longer than a commit message.
Agentic Ship — the plain engine for shipping real products with agents: deterministic UI system, wired Convex backend seams, and verification gates, delivered as skills.
The Agentic Ship agentic product system as an installable plugin: durable product contracts, neutral specialist roles, resumable service connections, deterministic UI and backend procedures, security, testing, health, and launch gates.
Gives the agent up-to-date documentation and code examples for libraries and frameworks, fetched on demand so answers are not stuck at the model's training cutoff. Runs locally from the @upstash/context7-mcp npm package.