Forge website

Skill ForgeyClap/claude-forge/.claude/skills/forge-website

Zero-dependency multi-agent build, automation and review system for Claude Code. 18 agents, 23 skills, a live per-project dashboard, and a /setup-forge onboarding wizard. One command: /forge.

Install

npx -y skills add ForgeyClap/claude-forge --skill forge-website

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

2 things to look at

18 days oldThe repository was created 18 days ago. New is not bad, but a brand new repository carrying a familiar-sounding name is the shape a typosquat arrives in, and there has been no time for anyone else to find a problem with it.
2 stars2 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Forge playbook for websites, landing pages, and frontend UI. Use when building or improving a website, landing page, marketing site, dashboard UI, or frontend — keywords: website, landing page, hero, CTA, responsive, mobile layout, UI, UX, accessibility, a11y, SEO, page speed, Lighthouse, frontend, React, Vue, Tailwind. Covers UI/UX, responsiveness, accessibility, SEO, performance, forms, and screenshot-loop review.

SKILL.md

2.7 KB, as published. Nobody here has run it

Forge playbook — Website / landing / frontend

Do not duplicate ECC skills — defer to: design-is (design audit), browser (screenshot loop). This file is orchestration only.

Hard rules

Mobile and desktop layout verified by screenshot before claiming done.
No placeholder/Lorem/test/mock content in anything headed for prod.
Forms validate input and hit a real endpoint; clear error + success states.
Any analytics/form keys live in env, never in the committed bundle.

Team (conditional by stack/level)

Lead: planner / architect.
Build review: react-reviewer or vue-reviewer (by stack).
Audits (parallel): a11y-architect, seo-specialist, performance-optimizer.
Copy/CTA: marketing-agent.

Skills / commands / MCP

design-is (pick a real style direction first — see ~/.claude/rules/ecc/web/design-quality.md), browser (mobile+desktop screenshots), /react-build + /react-review (or the Vue equivalents), /test-coverage. Canva MCP only if asset generation is explicitly requested. humanizer skill (.claude/skills/humanizer/SKILL.md) MAY run as an on-demand rewrite pass on client-facing marketing/landing copy — never on exact-command setup docs or forge-report output, and not always-on (34KB context cost). When a build uses GSAP/ScrollTrigger for scroll-driven or compositor-friendly motion (per ~/.claude/rules/ecc/web/performance.md), load the relevant gsap skill (.claude/skills/gsap/<skill-name>/SKILL.md) on demand for correct API usage — not always-on.

Fan-out & flow

L2 for a single page, L3 for a multi-page site. Serial: design direction (design-is) → build → audits → fixes → review gate. Parallel: a11y ∥ seo ∥ performance ∥ copy audits on the built artifact (independent).

Domain gates

Accessibility pass; SEO meta/OG + canonical; performance budget (see ~/.claude/rules/ecc/web/performance.md); responsive at ~360/768/1280; CTA clarity; no broken links.

Ship-readiness (unique)

Screenshots at mobile+desktop breakpoints; no test copy; all CTAs/links resolve; forms submit to a real endpoint with validation; favicon/meta/OG present; 404 route; secrets in env. The ship-readiness website checklist is advisory; optionally run codex-reviewer (Codex) on important code — not a blocker.