start-vibing 4.4.12 → 4.4.14

package/package.json

@@ -1,6 +1,6 @@
 {
 	"name": "start-vibing",
-	"version": "4.4.12",
+	"version": "4.4.14",
 	"description": "Setup Claude Code with 9 plugins, 6 community skills, and 8 MCP servers. Parallel install, auto-accept, superpowers + ralph-loop. e2e-audit 0.2.0 refactor (skill-only, no agents): SessionStart hook + slash command make the skill keyword-invokable (\"e2e audit\", \"roda o e2e\", \"integration test\", \"test coverage gaps\"). Source-first discovery via detect-stack, discover-routes (Next app/pages/Remix/SvelteKit/Nuxt/Astro), discover-api-surface (HTTP handlers, tRPC procedures, GraphQL, server actions, middleware auth), inventory-existing-tests (preserve prior corpus + sha256 drift hash), and detect-uncovered (branch-diff vs origin/main finds changes not covered by existing specs). Report-then-ask between mapping and Playwright run; post-run-feedback report before writing findings. SHOT+TRACE+ASSERT+SOURCE evidence quad per non-meta finding; meta rules (coverage-gap-*, uncovered-*, test-drift, stack-detect, post-run-feedback) exempt. verify-audit.sh enforces schema + quad. Generic (no project leakage). super-design 0.7.0 carries over.",
 	"type": "module",
 	"bin": {

package/template/.claude/CLAUDE.md

@@ -5,16 +5,34 @@ For user-facing project rules, see `/CLAUDE.md`.
 
 ---
 
+## Language Policy (HARD RULE)
+
+> **ENGLISH ONLY — for everything Claude produces — UNLESS the user explicitly asks for another language in the current request.**
+
+Applies to **every** output, no exceptions beyond an explicit user override:
+
+- Chat replies, status updates, summaries, questions back to the user
+- Code (identifiers, variables, function/class names, type names)
+- Comments, JSDoc, docstrings
+- Commit messages, PR titles, PR descriptions, branch names
+- Documentation (`/docs/**`, README, CLAUDE.md updates, changelogs, ADRs)
+- Test names, error messages, log messages, CLI help strings
+- File and folder names, configuration keys
+
+The user may write to Claude in Portuguese, Spanish, or any other language — Claude still responds and writes in English unless the user explicitly says e.g. "responda em português", "write this in Spanish". An override applies only to the current request; revert to English on the next turn. Never mix English + another language in the same output. If unsure whether a request is an override, default to English and ask one short clarifying question.
+
+---
+
 ## What start-vibing v4 Installs
 
 start-vibing is a CLI (`npx start-vibing`) that sets up Claude Code with a complete development system in ~30 seconds:
 
-| Component | Count | Installation Method |
-|-----------|-------|-------------------|
-| **MCP Servers** | 8 | `claude mcp add` (parallel, ~20s) |
-| **Plugins** | 9 | `claude plugin install` (parallel, auto-accept, ~3s) |
-| **Community Skills** | 5 | Direct HTTPS download from GitHub (~0.3s) |
-| **Template Files** | ~30 | File copy (agents, skills, config, CLAUDE.md) |
+| Component            | Count | Installation Method                                  |
+| -------------------- | ----- | ---------------------------------------------------- |
+| **MCP Servers**      | 8     | `claude mcp add` (parallel, ~20s)                    |
+| **Plugins**          | 9     | `claude plugin install` (parallel, auto-accept, ~3s) |
+| **Community Skills** | 5     | Direct HTTPS download from GitHub (~0.3s)            |
+| **Template Files**   | ~30   | File copy (agents, skills, config, CLAUDE.md)        |
 
 ### Installation Architecture
 
@@ -32,14 +50,14 @@ All phases run best-effort. MCP and plugin failures are non-blocking — `settin
 
 ### CLI Options
 
-| Flag | Effect |
-|------|--------|
-| `--force` | Overwrite all files (including custom ones) |
-| `--new` | Start fresh Claude session (default: resume last) |
-| `--no-claude` | Skip Claude Code installation and launch |
-| `--no-mcp` | Skip MCP server installation |
-| `--no-skills` | Skip community skills installation |
-| `--no-update-check` | Skip version check |
+| Flag                | Effect                                            |
+| ------------------- | ------------------------------------------------- |
+| `--force`           | Overwrite all files (including custom ones)       |
+| `--new`             | Start fresh Claude session (default: resume last) |
+| `--no-claude`       | Skip Claude Code installation and launch          |
+| `--no-mcp`          | Skip MCP server installation                      |
+| `--no-skills`       | Skip community skills installation                |
+| `--no-update-check` | Skip version check                                |
 
 ---
 
@@ -58,13 +76,13 @@ All phases run best-effort. MCP and plugin failures are non-blocking — `settin
 
 ## Agents (5 Active Subagents)
 
-| Agent | Model | Purpose |
-|-------|-------|---------|
-| **research-web** | sonnet | Researches best practices (2025-2026) with ontology-based structuring, output to `/docs/research/` |
-| **documenter** | sonnet | Analyzes sessions via git log/diff, writes changelog + technical docs + ADRs to `/docs/` |
-| **commit-manager** | haiku | Manages git commits, conventional format, merge workflow |
-| **claude-md-compactor** | sonnet | Compacts CLAUDE.md when it exceeds 40k chars |
-| **tester-unit** | sonnet | Creates unit tests with Vitest for new functions and utilities |
+| Agent                   | Model  | Purpose                                                                                            |
+| ----------------------- | ------ | -------------------------------------------------------------------------------------------------- |
+| **research-web**        | sonnet | Researches best practices (2025-2026) with ontology-based structuring, output to `/docs/research/` |
+| **documenter**          | sonnet | Analyzes sessions via git log/diff, writes changelog + technical docs + ADRs to `/docs/`           |
+| **commit-manager**      | haiku  | Manages git commits, conventional format, merge workflow                                           |
+| **claude-md-compactor** | sonnet | Compacts CLAUDE.md when it exceeds 40k chars                                                       |
+| **tester-unit**         | sonnet | Creates unit tests with Vitest for new functions and utilities                                     |
 
 ### Agent Workflow Order
 
@@ -79,6 +97,7 @@ Agents are dispatched via the `Agent` tool with `subagent_type` matching agent n
 The research-web agent outputs findings to `/docs/research/[topic].md` (NOT to `.claude/skills/research-cache/`).
 
 **Research flow:**
+
 1. Check `/docs/research/` for existing findings (reuse if fresh < 3 months)
 2. Build ontology map (concepts → relationships → constraints)
 3. Search with `[topic] [aspect] [2025-2026] [context]` queries
@@ -86,6 +105,7 @@ The research-web agent outputs findings to `/docs/research/[topic].md` (NOT to `
 5. Save structured output to `/docs/research/`
 
 **For UI/UX tasks, the agent also runs:**
+
 - Competitor analysis (3-5 competitors, heuristic evaluation)
 - Design system pattern check (shadcn/ui, Radix, WCAG 2.1)
 - User flow mapping (happy path + 2 error paths)
@@ -96,6 +116,7 @@ The research-web agent outputs findings to `/docs/research/[topic].md` (NOT to `
 The documenter agent runs after implementation, analyzes the session, and writes structured docs to `/docs/`.
 
 **Documentation flow:**
+
 1. Run `git log` + `git diff` to analyze what changed
 2. Classify changes: per-commit, per-feature, per-session
 3. Mini-research for technologies used (check `/docs/research/` first, else 1-2 queries)
@@ -103,6 +124,7 @@ The documenter agent runs after implementation, analyzes the session, and writes
 5. Update all indexes (`/docs/index.md` + per-folder indexes)
 
 **Output structure:**
+
 ```
 /docs/
 ├── index.md              # Root index
@@ -121,6 +143,7 @@ The documenter agent runs after implementation, analyzes the session, and writes
 ```
 
 **Writing rules:**
+
 - Self-contained sections (AI RAG chunk retrieval)
 - What→Why→How progression (humans first)
 - Before→After pattern for all changes (mandatory)
@@ -136,22 +159,22 @@ Plugins are the primary extension mechanism. All 9 are installed in parallel wit
 
 ### Core Workflow Plugins
 
-| Plugin | Mechanism | Purpose |
-|--------|-----------|---------|
-| **superpowers** | Skills + commands | TDD, brainstorming, debugging, planning, code review, git worktrees |
-| **ralph-loop** | Stop hook + command | Iterative autonomous development loop with checkpoints |
-| **context7** | Skill + agent + MCP | Auto library docs — replaces manual context7 MCP server |
-| **code-simplifier** | Skill + command | Refine code for clarity, reduce nesting, improve naming |
+| Plugin              | Mechanism           | Purpose                                                             |
+| ------------------- | ------------------- | ------------------------------------------------------------------- |
+| **superpowers**     | Skills + commands   | TDD, brainstorming, debugging, planning, code review, git worktrees |
+| **ralph-loop**      | Stop hook + command | Iterative autonomous development loop with checkpoints              |
+| **context7**        | Skill + agent + MCP | Auto library docs — replaces manual context7 MCP server             |
+| **code-simplifier** | Skill + command     | Refine code for clarity, reduce nesting, improve naming             |
 
 ### Development Plugins
 
-| Plugin | Mechanism | Purpose |
-|--------|-----------|---------|
-| **typescript-lsp** | LSP server (auto) | Type diagnostics, go-to-definition, hover info |
-| **security-guidance** | PreToolUse hook (auto) | OWASP Top 10, vulnerability scan, blocks unsafe patterns |
-| **code-review** | `/code-review` command | Code quality analysis, PR review |
-| **commit-commands** | `/commit` command | Git commit, push, PR flows with conventional format |
-| **frontend-design** | `/frontend-design` command | Production-grade UI design with competitor research |
+| Plugin                | Mechanism                  | Purpose                                                  |
+| --------------------- | -------------------------- | -------------------------------------------------------- |
+| **typescript-lsp**    | LSP server (auto)          | Type diagnostics, go-to-definition, hover info           |
+| **security-guidance** | PreToolUse hook (auto)     | OWASP Top 10, vulnerability scan, blocks unsafe patterns |
+| **code-review**       | `/code-review` command     | Code quality analysis, PR review                         |
+| **commit-commands**   | `/commit` command          | Git commit, push, PR flows with conventional format      |
+| **frontend-design**   | `/frontend-design` command | Production-grade UI design with competitor research      |
 
 ### Plugin Installation Details
 
@@ -193,6 +216,7 @@ All file changes and git history persist between iterations. Best for multi-file
 ### Using Code Simplifier (POST-IMPLEMENTATION)
 
 After implementing, run `/simplify` to:
+
 - Reduce nesting and redundancy
 - Improve naming and readability
 - Replace nested ternaries with early returns
@@ -206,10 +230,10 @@ All 8 MCPs are installed in parallel via `claude mcp add -s user` (~20s total).
 
 ### Required MCPs (ALWAYS use these)
 
-| Server | Package | Purpose |
-|--------|---------|---------|
+| Server                | Package                                            | Purpose                                                                      |
+| --------------------- | -------------------------------------------------- | ---------------------------------------------------------------------------- |
 | `sequential-thinking` | `@modelcontextprotocol/server-sequential-thinking` | Structured reasoning for multi-step tasks, architecture decisions, debugging |
-| `playwright` | `@playwright/mcp@latest` | Browser automation for UI verification, E2E tests, visual validation |
+| `playwright`          | `@playwright/mcp@latest`                           | Browser automation for UI verification, E2E tests, visual validation         |
 
 > These 2 are **non-negotiable**. Skipping them leads to poor planning and untested UIs.
 
@@ -217,27 +241,27 @@ All 8 MCPs are installed in parallel via `claude mcp add -s user` (~20s total).
 
 ### Standard MCPs
 
-| Server | Package | Transport | Purpose |
-|--------|---------|-----------|---------|
-| `memory` | `@modelcontextprotocol/server-memory` | stdio (npx) | Persistent knowledge graph across sessions |
-| `nextjs-devtools` | `next-devtools-mcp@latest` | stdio (npx) | Next.js runtime errors, routes, cache inspection |
-| `mongodb` | `@mongodb-js/mongodb-mcp-server` | stdio (npx) | MongoDB queries, schema inspection, aggregation |
-| `jira` | `@aashari/mcp-server-atlassian-jira` | stdio (npx) | Issue tracking, task management |
-| `git` | `mcp-server-git` | stdio (uvx) | Git operations, repo search, history, diffs |
-| `fetch` | `mcp-server-fetch` | stdio (uvx) | Fetch web pages as markdown |
+| Server            | Package                               | Transport   | Purpose                                          |
+| ----------------- | ------------------------------------- | ----------- | ------------------------------------------------ |
+| `memory`          | `@modelcontextprotocol/server-memory` | stdio (npx) | Persistent knowledge graph across sessions       |
+| `nextjs-devtools` | `next-devtools-mcp@latest`            | stdio (npx) | Next.js runtime errors, routes, cache inspection |
+| `mongodb`         | `@mongodb-js/mongodb-mcp-server`      | stdio (npx) | MongoDB queries, schema inspection, aggregation  |
+| `jira`            | `@aashari/mcp-server-atlassian-jira`  | stdio (npx) | Issue tracking, task management                  |
+| `git`             | `mcp-server-git`                      | stdio (uvx) | Git operations, repo search, history, diffs      |
+| `fetch`           | `mcp-server-fetch`                    | stdio (uvx) | Fetch web pages as markdown                      |
 
 ### Optional MCPs (install manually)
 
 These are shown to the user after setup but not auto-installed:
 
-| Server | Install Command |
-|--------|----------------|
+| Server   | Install Command                                                                     |
+| -------- | ----------------------------------------------------------------------------------- |
 | `github` | `claude mcp add --transport http -s user github https://api.githubcopilot.com/mcp/` |
-| `sentry` | `claude mcp add --transport http -s user sentry https://mcp.sentry.dev/mcp` |
-| `figma` | `claude mcp add --transport http -s user figma https://mcp.figma.com/mcp` |
-| `linear` | `claude mcp add --transport http -s user linear https://mcp.linear.app/sse` |
-| `stripe` | `claude mcp add --transport http -s user stripe https://mcp.stripe.com` |
-| `vercel` | `claude mcp add --transport http -s user vercel https://mcp.vercel.com` |
+| `sentry` | `claude mcp add --transport http -s user sentry https://mcp.sentry.dev/mcp`         |
+| `figma`  | `claude mcp add --transport http -s user figma https://mcp.figma.com/mcp`           |
+| `linear` | `claude mcp add --transport http -s user linear https://mcp.linear.app/sse`         |
+| `stripe` | `claude mcp add --transport http -s user stripe https://mcp.stripe.com`             |
+| `vercel` | `claude mcp add --transport http -s user vercel https://mcp.vercel.com`             |
 
 ---
 
@@ -260,13 +284,13 @@ Community skills are downloaded directly from GitHub raw URLs (the `skillsadd` n
 
 ### Installed Skills
 
-| Skill | Source Repo | Purpose |
-|-------|------------|---------|
-| **react-best-practices** | `vercel-labs/agent-skills` | React/Next.js performance optimization rules |
-| **web-design-guidelines** | `vercel-labs/agent-skills` | 100+ WCAG accessibility + UX audit rules |
-| **composition-patterns** | `vercel-labs/agent-skills` | Compound component and composition patterns |
-| **webapp-testing** | `anthropics/skills` | Real browser test execution with Playwright |
-| **mcp-builder** | `anthropics/skills` | Guide for building MCP servers |
+| Skill                     | Source Repo                | Purpose                                      |
+| ------------------------- | -------------------------- | -------------------------------------------- |
+| **react-best-practices**  | `vercel-labs/agent-skills` | React/Next.js performance optimization rules |
+| **web-design-guidelines** | `vercel-labs/agent-skills` | 100+ WCAG accessibility + UX audit rules     |
+| **composition-patterns**  | `vercel-labs/agent-skills` | Compound component and composition patterns  |
+| **webapp-testing**        | `anthropics/skills`        | Real browser test execution with Playwright  |
+| **mcp-builder**           | `anthropics/skills`        | Guide for building MCP servers               |
 
 ### Adding More Skills
 
@@ -294,12 +318,12 @@ curl -o .claude/skills/<name>/SKILL.md \
 
 Project-specific settings in `.claude/config/`:
 
-| File | Purpose |
-|------|---------|
-| `project-config.json` | Stack, structure, commands |
-| `quality-gates.json` | Quality check commands |
+| File                  | Purpose                        |
+| --------------------- | ------------------------------ |
+| `project-config.json` | Stack, structure, commands     |
+| `quality-gates.json`  | Quality check commands         |
 | `testing-config.json` | Test framework and conventions |
-| `security-rules.json` | Security audit rules |
+| `security-rules.json` | Security audit rules           |
 
 Agents read config files before acting. Do NOT hardcode project specifics — update the JSON files instead.
 
@@ -311,17 +335,17 @@ The `.claude/settings.json` file contains `enabledPlugins` which is the fallback
 
 ```json
 {
-  "enabledPlugins": {
-    "typescript-lsp@claude-plugins-official": true,
-    "security-guidance@claude-plugins-official": true,
-    "code-review@claude-plugins-official": true,
-    "commit-commands@claude-plugins-official": true,
-    "frontend-design@claude-plugins-official": true,
-    "superpowers@claude-plugins-official": true,
-    "ralph-loop@claude-plugins-official": true,
-    "context7@claude-plugins-official": true,
-    "code-simplifier@claude-plugins-official": true
-  }
+	"enabledPlugins": {
+		"typescript-lsp@claude-plugins-official": true,
+		"security-guidance@claude-plugins-official": true,
+		"code-review@claude-plugins-official": true,
+		"commit-commands@claude-plugins-official": true,
+		"frontend-design@claude-plugins-official": true,
+		"superpowers@claude-plugins-official": true,
+		"ralph-loop@claude-plugins-official": true,
+		"context7@claude-plugins-official": true,
+		"code-simplifier@claude-plugins-official": true
+	}
 }
 ```
 
@@ -359,6 +383,7 @@ The `.claude/settings.json` file contains `enabledPlugins` which is the fallback
 ### After Completing Work
 
 The documenter agent runs automatically after implementation:
+
 1. Analyzes git log/diff for the session
 2. Writes changelog + technical docs + ADRs as needed
 3. Updates all indexes
@@ -397,23 +422,23 @@ All implementations MUST:
 
 ## FORBIDDEN Actions
 
-| Action                         | Reason                       |
-| ------------------------------ | ---------------------------- |
-| Write in non-English           | ALL code/docs MUST be in EN  |
-| Skip typecheck                 | Catches runtime errors       |
-| Use `any` type                 | Defeats strict mode          |
-| Define types in `src/`         | Must be in `types/`          |
-| Commit directly to main        | Create feature/fix branches  |
-| Skip documenter after implementation | Changelog + docs are mandatory |
-| Mix doc types in one file      | Changelog ≠ technical ≠ decision |
-| Leave docs unlinked from index | Undiscoverable docs are useless |
-| Skip superpowers for features  | Use brainstorming + TDD      |
-| Skip code-simplifier           | Run /simplify post-implementation |
-| Use MUI/Chakra                 | Use shadcn/ui + Radix        |
-| Files > 400 lines              | MUST split into smaller      |
-| 'use client' at top level      | Push to leaf components only |
-| Waterfall data fetching        | Use Promise.all() for parallel |
-| Skip CLAUDE.md update          | MUST update after implementations |
+| Action                               | Reason                                                                                                                |
+| ------------------------------------ | --------------------------------------------------------------------------------------------------------------------- |
+| Speak/write in non-English           | EN ONLY (chat, code, docs, commits) — see Language Policy. Override only via explicit user request, current turn only |
+| Skip typecheck                       | Catches runtime errors                                                                                                |
+| Use `any` type                       | Defeats strict mode                                                                                                   |
+| Define types in `src/`               | Must be in `types/`                                                                                                   |
+| Commit directly to main              | Create feature/fix branches                                                                                           |
+| Skip documenter after implementation | Changelog + docs are mandatory                                                                                        |
+| Mix doc types in one file            | Changelog ≠ technical ≠ decision                                                                                      |
+| Leave docs unlinked from index       | Undiscoverable docs are useless                                                                                       |
+| Skip superpowers for features        | Use brainstorming + TDD                                                                                               |
+| Skip code-simplifier                 | Run /simplify post-implementation                                                                                     |
+| Use MUI/Chakra                       | Use shadcn/ui + Radix                                                                                                 |
+| Files > 400 lines                    | MUST split into smaller                                                                                               |
+| 'use client' at top level            | Push to leaf components only                                                                                          |
+| Waterfall data fetching              | Use Promise.all() for parallel                                                                                        |
+| Skip CLAUDE.md update                | MUST update after implementations                                                                                     |
 
 ---
 

package/template/.claude/skills/typeui-web3-lp/SKILL.md

@@ -0,0 +1,407 @@
+---
+name: web3-lp
+description: Cyber-tactical landing-page system for Web3 token launches with real product behind them — sticky live-price bar, hex-grid scanline hero, animated track-record proof, floating buy pill, "WHAT IS $TICKER" reveal. Built for fair launches (pump.fun / fair-launch DEX) where the coin is fun but the tech is real. Use when user says "build a token LP", "coin launch page", "web3 LP", "memecoin site with tech", "$TICKER landing", "pump.fun launch site".
+license: MIT
+metadata:
+    author: deadodds-build
+    origin: 'Synthesized from sd-research audits of HyperLiquid, Pump.fun, dYdX, GMX, Mog, Brett, Ondo, Polymarket, Limitless, Drift, Kalshi, Birdeye, Jupiter — and the DeadOdds AI / $ODDS landing page (2026-04-28).'
+---
+
+# Web3-Launch Design System Skill (Universal)
+
+## Mission
+
+You are an expert design-system author for **Web3 token-launch landing pages** — coin LPs that have real tech (not pure memes) and want both degen credibility AND product credibility.
+
+The single hardest thing about a token-with-tech LP is balancing two audiences: degens scanning for ticker + chart + buy-button in <3 seconds, and serious users who need to verify the product works. This skill exists to resolve that tension.
+
+When unsure: prioritize **proof** over **polish**. Numbers > adjectives. Live data > marketing copy. Receipts > promises.
+
+## Brand North-Star
+
+> **The token is fun. The product is real.**
+
+Every section either hypes the coin OR proves the tech. Never both at once. Every claim has a number behind it. Every button has a destination. Nothing is decorative — the cyberpunk aesthetic is the brand voice, not a costume.
+
+## Style Foundations
+
+### Visual style
+
+- Dark base (`#050505` to `#0A0A0A`), single neon accent, monospace data, geometric display headlines
+- Hex-grid or terminal-scanline hero background (open territory; competitors use blob gradients or photography)
+- Glow-shadow CTAs (subtle 28px spread, 4px offset) — never flat
+- Sharp corners (4–8px radius); never rounded-full except for floating pills
+
+### Typography scale
+
+- **Display:** Rajdhani / Orbitron / Geist condensed bold, ALL CAPS, tight tracking
+- **Body:** Inter / Geist Sans, sentence case
+- **Numbers / data:** Geist Mono / JetBrains Mono / IBM Plex Mono, tabular-nums
+- **Scale:** 10/11/12/13.5/14.5/16/22/24/32/48/clamp-hero
+
+### Color palette (default — override if user has tokens)
+
+- `--brand-bg: #050505` (near-black surface)
+- `--brand-surface: #0A0A0A` (cards)
+- `--brand-accent: #7CFF3D` (neon — primary CTA + win signal)
+- `--brand-danger: #FF3D5C` (lose signal / warning)
+- `--brand-chrome: #C7CDD4` (secondary text)
+- `--brand-text: #FFFFFF` (primary text)
+- `--brand-text-muted: #64748B` (slate-500 — labels, captions)
+- `--brand-line: #1A1A1A` (1px borders)
+
+If the user has a token color (memecoin tradition: $MOG = gold, $BRETT = cyan, $PEPE = lime), keep `#7CFF3D` semantics for **GREEN/win** and use the user's color as a **secondary brand accent** for hero glow + token block emphasis. Never use 3+ accents.
+
+### Spacing scale
+
+- `4 / 8 / 12 / 16 / 20 / 24 / 32 / 48 / 64 / 80 / 112 / 144`
+- Section vertical rhythm: `py-20 lg:py-28` minimum between major blocks
+- Cards: `p-5 lg:p-7`
+- Compact UI (price bar, tickers): `py-2.5 px-4`
+
+### Border radius
+
+- Buttons / inputs / cards / pills: `rounded` (4px) — sharp, technical
+- Floating buy pill: `rounded-full` — only exception
+- Token block accent boxes: `rounded` (4–6px)
+- NEVER `rounded-2xl` / `rounded-3xl` — that's SaaS, not Web3-tactical
+
+## Section Playbook (the LP recipe)
+
+This is the canonical section order for a Web3-launch LP. Skip sections you don't have data for, but DO NOT reorder. The flow has been triangulated across 12 audited competitors.
+
+```
+01. PriceBar         — sticky top, h-9, live ticker + MCap + Vol + buy links
+02. MarketingHeader  — sticky top-9, wordmark + nav + social count badges + Launch App
+03. LiveTicker       — marquee of recent product events (signals / trades / mints)
+04. Hero             — hex-grid bg, headline, dual CTAs, contract row, in-hero proof row
+05. WhatIs           — full-screen reveal: "WHAT IS $TICKER?" with poetic 2-liner
+06. TrackRecord      — animated counters + sample resolved/historical data table
+07. HowItWorks       — 3 steps, numbered cards (01 / 02 / 03), big oversized digit watermark
+08. TechPreview      — "Not a meme. A weapon." — mock product UI screenshot or live embed
+09. TokenBlock       — contract card + tokenomics 2x2 grid + buy CTAs
+10. Community        — 3-card grid: Telegram / X / pump.fun (or chain-native exchange)
+11. FAQ              — <details> accordion, 5–7 questions, first item open
+12. FinalCta         — "stop guessing, start sniping" pattern
+13. MarketingFooter  — 4 columns (brand / Product / Token / Community)
+14. FloatingBuy      — fixed bottom-right pill, appears after hero scroll
+```
+
+### Section anatomy rules
+
+**01. PriceBar** (mandatory if token is launched; mock with `TBA` placeholders pre-launch)
+
+- Height `h-9` (36px), `sticky top-0 z-[60]`, `bg-black/95 backdrop-blur-md`
+- Left side: `· $TICKER $price ±change% | MCap $X | Vol 24h $Y | Holders N`
+- Right side: `pump.fun ↗ | Dexscreener ↗`
+- Pulsing dot before ticker for "live" signal (`animate-pulse` + accent glow)
+- Hide non-essential cells progressively at sm/md/lg breakpoints
+- ALWAYS connectable to the launchpad's public API once token is live
+
+**02. MarketingHeader**
+
+- `sticky top-9 z-50 h-14 lg:h-16`
+- Wordmark left, primary nav center, social-count badges + Launch App right
+- Social badges show actual follower counts (e.g. `4,221`) on desktop
+- Primary CTA = "Launch App" → links to the actual product, NOT to buy page
+- Buy CTA lives in PriceBar + FloatingBuy, never in the header
+
+**03. LiveTicker**
+
+- 60s loop marquee, full-width, `border-y border-line bg-surface`
+- Items show ACTUAL product events, not slogans (e.g. "AI: GREEN +14.2pp on [Trump 2028]" not "BUY $ODDS NOW")
+- Pause on hover (already in `globals.css`)
+
+**04. Hero** (the most important section)
+
+- Hex-grid + scanline background — see Motion Catalog
+- Eyebrow: "Fair-launching $TICKER on pump.fun" with pulsing dot
+- Headline: 2 lines, second line is the brand color with `textShadow: 0 0 28px <accent>45`
+- Add terminal-cursor blink (`_`) at end of color line — code-aesthetic signal
+- Sub-copy: 2 sentences max, bold the receipts ("EV-edge bets", "real time", "with receipts")
+- Dual CTA: Primary = "Buy $TICKER on pump.fun" (filled accent) · Secondary = "Launch Oracle/App" (ghost, `bg-black/50 backdrop-blur`)
+- ContractRow card directly below CTAs — NEVER bury contract address in token section
+- In-hero proof row (3 stats max) below contract — animated counters that fire on first paint
+- Stats accent: middle stat is the headline metric (win rate / TVL / users) in brand color
+
+**05. WhatIs** (the moment of poetry)
+
+- `min-h-[80vh]`, oversized type at `clamp(3rem, 11vw, 9rem)`
+- Eyebrow as terminal command: `> query: what_is_$ticker`
+- Headline structure: "WHAT IS" newline "$TICKER?" — second line in accent
+- 2–3 short statements (≤8 words each), the third in muted color as denouement
+- Two CTAs: "Buy $TICKER" + "See the receipts" (anchor to TrackRecord)
+
+**06. TrackRecord** (the onliness section — what no competitor has)
+
+- 4-stat grid (counters animate on scroll into view)
+- Resolved-events / sample-data table: desktop = `<table>`, mobile = card list (NEVER a horizontal scroll)
+- Header row: monospace eyebrow `▸ resolved_predictions · sample` + Brier score / best result on the right
+- Each row: market name + AI/your call + market consensus + resolution badge + edge captured
+- Footer link: "See live signal feed in the [product] ›"
+
+**07. HowItWorks**
+
+- 3 numbered cards in a `lg:grid-cols-3`
+- Big watermark digit (`01 / 02 / 03`) at top-right, opacity `0.025` → `0.10` on hover
+- Card eyebrow: "Step 01" in accent monospace
+- Card headline: ALL CAPS display, ≤3 words
+- Card body: 2 sentences max, plain language
+
+**08. TechPreview**
+
+- Two-column grid `lg:grid-cols-2 lg:items-center`
+- Left = copy + bullet list + CTA · Right = mock product UI in faux window chrome
+- Mock window chrome: 3 dots + tab label `<wordmark> · live signals`
+- 3 mock cards inside, one each: positive / negative / neutral signal — colored accent border
+- Headline pattern: "Not a meme. A weapon." (or domain equivalent — "Not a fork. A protocol.")
+
+**09. TokenBlock**
+
+- 2-column `lg:grid-cols-[1.1fr_1fr]`: hero card on left, tokenomics 2×2 on right
+- Hero card has accent radial blur top-right + ticker huge + chain badge + contract code-block + dual buy CTA
+- Tokenomics tiles: Supply / Launch / Liquidity / Tax — each with label + big number + 1-line caption
+- Default copy for fair launch: "Fair · 100% on pump.fun" / "Burned · LP locked at bond" / "0 / 0 · forever"
+
+**10. Community**
+
+- 3-card grid, each card = social platform with: brand-color icon, name, handle, blurb
+- Cards have subtle hover glow + arrow shifts right
+- Order: Telegram (community) · X (alpha) · pump.fun/Dexscreener (token)
+
+**11. FAQ**
+
+- Native `<details>` accordion (zero JS) inside `<dl>` semantics
+- First item `open` by default
+- "+" → "×" rotation on open via `group-open:rotate-45`
+- 5–7 questions; first MUST be "Is $TICKER required to use the app?" (or domain analog)
+
+**12. FinalCta**
+
+- Centered, `py-24 lg:py-32`, radial-gradient floor accent
+- Headline 2 lines, second line in brand color
+- Same dual CTAs as hero
+
+**13. Footer**
+
+- 4 columns: brand+blurb / Product / Token / Community
+- Disclaimer line: "Not financial advice. Markets are risky. Bet responsibly."
+- Right-aligned tagline: "Built on [chain] · Powered by [tech]"
+
+**14. FloatingBuy** (the conversion safety net)
+
+- Fixed `bottom-5 right-4 lg:bottom-7 lg:right-7`
+- Appears after `scrollY > 600`, hides when `nearFooter` (avoid double-CTA collision)
+- `rounded-full` (only exception to sharp-radius rule), accent-filled
+- Scale-up + glow-up on hover (`hover:scale-[1.03]`)
+
+## Motion Catalog
+
+Six approved motion patterns. **Use ALL of them — collectively they make the LP feel alive, individually they feel decorative. Skip any and the page reads dead.**
+
+1. **Hex-grid scanline** — hero background. CSS gradient hex pattern + radial mask + 1 horizontal scanline sweeping top→bottom over 7s. Open territory; no audited competitor uses it.
+2. **Marquee tickers** — `LiveTicker` (60s loop), pause on hover, `animation-play-state: paused` on hover.
+3. **Animated counters** — IntersectionObserver fires when stat enters viewport at 0.4 threshold; cubic-eased count-up over 1.4s; honors `prefers-reduced-motion`.
+4. **Terminal cursor blink** — single `_` char at end of hero accent line; 1.05s step-end. Reads as "command line", not decorative.
+5. **Floating-pill entrance** — translate-y(20px) + scale(0.9) → identity, 240ms ease-out. Plays each time the pill mounts.
+6. **CTA glow on hover** — primary CTAs gain `box-shadow: 0 0 40px -2px var(--accent)85` on hover, 150ms transition.
+
+**Do NOT add:** parallax (every Web3 LP has it; you'll look generic), full-page scroll-snapping, fade-in-on-scroll for every element (uncanny / slow), confetti, character mascots dancing.
+
+**MUST honor `prefers-reduced-motion: reduce`** — `animation: none` for marquee, scanline, counters, cursor blink. The page must still be navigable and beautiful without motion.
+
+## Copy / Voice
+
+### Tone (Aaker × NN/g)
+
+- **Sincere:** -1 (we're not earnest, we're sharp)
+- **Excitement:** +2 (degen-adjacent, but with restraint)
+- **Competence:** +2 (technical, precise, monospace numbers)
+- **Sophistication:** +1 (cyberpunk minimalism, not luxury)
+- **Ruggedness:** +1 (sniper / weapon / edge metaphors are OK)
+
+### Vocabulary
+
+- **PREFER:** alpha · edge · receipts · ship · sniper · oracle · signal · degen · fair launch · proof · live · win rate · EV (expected value) · Kelly · in the wild
+- **AVOID:** revolutionary · disrupting · ecosystem · empower · journey · mission · synergy · best-in-class · world's first (unless you can prove it)
+- **NEVER:** "leverage", "robust", "seamless", "unlock the future"
+
+### Headline patterns
+
+- 2 lines, second is brand-accent
+- Imperative or noun phrase (avoid passive)
+- Examples that work: "AI sniper for / prediction markets" · "Stop guessing. / Start sniping." · "Not a meme. / A weapon." · "Receipts on chain. / Edge in the wild."
+- Examples that don't: "Welcome to the future of X" · "The next generation of Y"
+
+### CTA copy
+
+- **Primary buy:** "Buy $TICKER" or "Buy $TICKER on pump.fun"
+- **Primary product:** "Launch App" / "Launch Oracle" — never "Get started"
+- **Secondary:** "See the receipts" / "Read docs" / "See live signal feed"
+- Always pair with a `›` arrow glyph (not `→` not lucide arrow)
+
+### Semantic voice
+
+- All numbers tabular-nums always
+- Wins/positive: brand-accent color
+- Losses/negative: `--brand-danger`
+- Neutral/pending: amber / slate-300
+- Status pills: `border-color/40 bg-color/10 text-color`
+
+## Reference LPs Audited
+
+This skill was synthesized from primary research on these sites (full evidence in any sd-research session that targets `web3-launch`):
+
+| Tier                     | Sites                             | What we stole                                                                      |
+| ------------------------ | --------------------------------- | ---------------------------------------------------------------------------------- |
+| Coin-with-tech (premium) | HyperLiquid, GMX, Ondo            | In-hero proof row · animated stat counters · 3-stat hero                           |
+| Pure meme + culture      | Mog, Brett                        | Sticky price bar · floating buy pill · "WTF IS X?" reveal · contract at hero level |
+| Prediction markets       | Polymarket, Kalshi, Limitless     | Live activity ticker · category code (dark + green/red) · marquee above nav        |
+| Solana DeFi              | Jupiter, Drift, Birdeye, Pump.fun | Card-grid layouts · monospace data · neon glow shadows                             |
+
+**Open territory we own:** serious + expressive + terminal/radar aesthetic. No audited competitor uses hex-grid + scanline as hero background. No audited competitor surfaces verifiable on-chain win-rate stats in the LP.
+
+## Default Onliness Pattern
+
+Use this template for the LP's positioning statement:
+
+> "$TICKER is the only [category] token with [verifiable proof] — [what we have], [where the proof lives], [the cultural energy]."
+
+Example: "$ODDS is the only prediction-market token with a verifiable on-chain win rate — AI signals, proof on-chain, degen energy."
+
+## Component Families (LP-scoped)
+
+- price ticker bar
+- marquee live feed
+- hero with motion background
+- contract address row (with copy + explorer + buy)
+- animated stat counter
+- track-record table (desktop) + card list (mobile)
+- numbered step cards
+- mock product preview card
+- token contract card
+- tokenomics tile grid
+- social platform card
+- accordion FAQ
+- floating buy pill
+- 4-column footer
+- terminal-style status header (`▸ command_name`)
+
+## Accessibility (non-negotiable)
+
+- WCAG 2.2 AA contrast on all text (verify accent on dark = 7.0+ for body sizes)
+- Visible focus rings: `focus-visible:outline-2 outline-offset-2 outline-[var(--brand-accent)]`
+- Touch targets `>= 44px` mobile (price bar links exempt; buy CTAs always)
+- `<details>` accordion is keyboard + screen-reader native — DON'T rebuild it as JS
+- All animations `prefers-reduced-motion: reduce` aware
+- Marquee tickers: `role="marquee"` + `aria-label` + pause-on-hover required
+- Contract row: `aria-label="Copy contract address"` on copy button; show `Copied` confirmation
+- Color is never the only signal — always pair with text label or icon (e.g. signal pill says "GREEN" not just green color)
+
+## Writing Tone (for the skill consumer)
+
+Confident, sharp, low-jargon. Short sentences. Active voice. Numbers > adjectives. Receipts > promises. Examples > descriptions.
+
+## Rules: Do
+
+- Lead with the number, not the claim ("73% win rate" before "AI is good")
+- One section = one job (hype OR proof, never both)
+- Use brand-accent only for: primary CTAs, win signals, pulsing live dots, key stats — anywhere else is brand dilution
+- Animate sparingly; every motion must serve recognition or recall
+- Surface the contract address in the hero — never hide it in a token section
+- Write CTAs as commands, not descriptions ("Buy $ODDS" not "Get $ODDS now")
+- Pair every screenshot with a caption that says what it is
+- Keep PriceBar + FloatingBuy as escape hatches so the user is never >300vh from a buy action
+- Pre-launch: use `TBA` placeholders inside the FINAL UI shell — don't omit the section. Users will see the structure they'll buy into
+
+## Rules: Don't
+
+- Don't use rounded-2xl/3xl on cards — that's SaaS, this is tactical
+- Don't write "revolutionary", "disrupt", "next-gen", "future of X"
+- Don't add parallax to the hero — it's the most overused Web3 LP move
+- Don't put 3+ CTAs in the hero — primary buy + secondary product, that's it
+- Don't decorate stats with glyphs/icons — the number IS the decoration
+- Don't make the FAQ a JS-controlled accordion — `<details>` is better
+- Don't ship without `prefers-reduced-motion` overrides on every animation
+- Don't translate the LP — degen audience reads English universally; translation dilutes voice
+- Don't add a "subscribe to newsletter" CTA — degen audiences reject email gates
+- Don't put more than 5 footer columns — 4 max (brand / Product / Token / Community)
+
+## Expected Behavior
+
+- Always restate the user's positioning in one onliness sentence before generating sections
+- Use the section playbook order; only skip sections, never reorder
+- Pre-launch: still ship all sections, fill with mock data shaped like the real data ("2,847 markets analyzed", "$1.42M MCap" — degen-realistic, not "Lorem ipsum")
+- For monorepos / multiple LPs (e.g. one per token), keep this skill as the shared brand floor and override `--brand-accent` per token
+- When a section has no data analog (e.g. a pure meme has no `TrackRecord`), substitute with the closest meme-tier proof (community stats, bridge/exchange listings, viral tweet count) — never delete the section silently
+
+## Guideline Authoring Workflow
+
+1. Restate the user's brand in one onliness sentence (use the template above).
+2. Define palette + type tokens; only override defaults if the user has source-of-truth tokens.
+3. Walk the Section Playbook in order; for each section, name the data the section will display and the CTA destination.
+4. List all 6 motion patterns and confirm `prefers-reduced-motion` overrides exist.
+5. Write the FAQ first (it forces clarity on what the project actually is) — then the rest of the copy.
+6. Run a "no decoration" pass: delete every UI element that isn't a CTA, a proof, or a navigation aid.
+
+## Required Output Structure
+
+When generating a Web3-launch LP, deliver:
+
+- One-line onliness statement
+- Token + brand tokens (palette, type, radius, spacing)
+- Per-section markup with: data shape, copy, CTA targets, motion notes
+- Mobile + desktop layout decisions per section
+- Accessibility checklist (focus rings, reduced-motion, touch targets, ARIA labels)
+- Anti-patterns explicitly avoided
+- QA checklist (see below)
+
+## QA Checklist
+
+- [ ] PriceBar shows live ticker + at least 2 of {MCap, Vol, Holders}; uses `TBA` if pre-launch
+- [ ] Hero contains: eyebrow + 2-line headline + sub-copy + 2 CTAs + contract row + 3-stat proof row — in that order
+- [ ] In-hero counters animate on first paint (or scroll into view), not stuck at 0
+- [ ] Contract row has: address (full or 4-char-truncated) + Copy button + Explorer link + Buy link
+- [ ] WhatIs section has min-h-80vh and oversized clamp(3rem, 11vw, 9rem) headline
+- [ ] TrackRecord stats animate on scroll-into-view via IntersectionObserver
+- [ ] TrackRecord has both desktop table AND mobile card list (NEVER horizontal scroll)
+- [ ] FloatingBuy appears after 600px scroll and disappears within 280px of footer
+- [ ] PriceBar (`top-0`) and Header (`top-9`) stack correctly with no overlap at any scroll
+- [ ] All marquees pause on hover via `[role="marquee"]:hover > div` selector
+- [ ] All animations have `@media (prefers-reduced-motion: reduce)` override
+- [ ] No `rounded-2xl` / `rounded-3xl` anywhere except FloatingBuy (intentional pill)
+- [ ] FAQ uses native `<details>` not JS accordion
+- [ ] Footer has 4 columns max, includes "Not financial advice" disclaimer
+- [ ] All numbers use `tabular-nums` and `font-mono` (Geist Mono / JetBrains Mono)
+- [ ] Brand accent used ONLY on: primary CTAs, win signals, live dots, eyebrow callouts, hero accent line
+- [ ] Touch targets >= 44px on mobile
+- [ ] No use of "revolutionary", "disrupt", "next-gen", "ecosystem", "empower"
+- [ ] If LP has multiple language audiences: ship English only — degen audience reads it universally
+
+## Component Rule Expectations
+
+For every component, define:
+
+- **Default state:** layout, tokens, copy
+- **Hover state:** glow-up if interactive, color shift if link
+- **Focus-visible:** outline ring with brand-accent
+- **Active state:** brief scale-down or color flash
+- **Loading / pre-launch:** show skeleton OR `TBA` placeholder inside final UI shell
+- **Error / empty:** monospace status header (`▸ no_data_yet`) + helpful next-action link
+- **Mobile collapse:** which cells/columns hide; which become card-stacks
+
+## Quality Gates
+
+- Every section must have a quantifiable success metric (CTA click rate / scroll depth / time-on-page) — if you can't name it, the section probably doesn't belong
+- Accent color usage audit: count occurrences. If `--brand-accent` appears more than 12 times in the rendered LP, you're over-using it
+- "No-JS pass": disable JS — the LP must still render readable, navigable, and convey the pitch. Only counters, marquees, and the floating-buy pill should degrade
+- Lighthouse: Performance ≥ 90, Accessibility ≥ 95, Best-Practices ≥ 95, SEO ≥ 95 on the LP
+
+## Migration Notes (when retrofitting an existing LP)
+
+1. Don't try to redesign in-place — fork the route into a route group like `(marketing)/` with its own layout
+2. Pull the dashboard / app out of `/` first; LP takes the root URL, app moves to `/dashboard` or `/app`
+3. Add PriceBar + FloatingBuy LAST, after sections render — they're the conversion safety net, not the structure
+4. The hex-grid scanline goes on the Hero only — don't bleed it into other sections (loses meaning)
+5. Replace any existing "fade up on scroll" decorative animations with the 6 approved motion patterns — quality > quantity
+ <!-- TYPEUI_SH_MANAGED_END -->