luna-agents 2.0.1

package/commands/ll-i18n.md

@@ -0,0 +1,74 @@
+---
+name: ll-i18n
+displayName: Luna Internationalization
+description: Set up i18n — extract strings, generate translation keys, configure i18n library, RTL support
+version: 1.0.0
+category: development
+agent: luna-task-executor
+parameters:
+  - name: locales
+    type: string
+    description: Comma-separated locale codes (en,he,ar,es,fr,de,ja,zh)
+    required: false
+    prompt: true
+  - name: scope
+    type: string
+    description: Project or feature scope
+    required: true
+    prompt: true
+workflow:
+  - scan_hardcoded_strings
+  - extract_translation_keys
+  - configure_i18n_library
+  - generate_locale_files
+  - add_rtl_support
+  - update_components
+  - generate_i18n_report
+output:
+  - .luna/{current-project}/i18n/
+  - .luna/{current-project}/i18n-report.md
+prerequisites: []
+---
+
+# Luna Internationalization
+
+Extract strings and set up complete i18n support.
+
+## What This Command Does
+
+1. **Scan** — finds all hardcoded user-facing strings in components
+2. **Extract** — generates translation keys with namespace organization
+3. **Configure** — sets up next-intl / react-i18next / vue-i18n
+4. **Generate** — creates locale JSON files with English defaults
+5. **RTL** — adds right-to-left support for Arabic, Hebrew
+6. **Update** — replaces hardcoded strings with `t('key')` calls
+7. **Report** — documents all extracted strings and configuration
+
+## Output Structure
+
+```
+locales/
+  en.json            # English (source)
+  he.json            # Hebrew (RTL)
+  ar.json            # Arabic (RTL)
+  es.json            # Spanish
+  fr.json            # French
+i18n.config.ts       # Library configuration
+middleware.ts        # Locale detection middleware
+```
+
+## Usage
+
+```
+/i18n en,he,ar              # English + Hebrew + Arabic
+/i18n en,es,fr,de,ja        # Multi-language
+```
+
+## Features
+
+- Namespace organization (common, auth, dashboard, errors)
+- Pluralization rules per locale
+- Date/number formatting per locale
+- RTL layout switching with `dir="rtl"`
+- Language switcher component
+- SEO: hreflang tags and locale-prefixed routes

package/commands/ll-idea.md

@@ -0,0 +1,101 @@
+---
+name: ll-idea
+displayName: Luna Idea → Product
+description: Transform a product idea into a full spec, architecture, design system, and implementation plan in one command
+version: 1.0.0
+category: creation
+agent: luna-design-architect
+parameters:
+  - name: idea
+    type: string
+    description: Your product idea in plain English
+    required: true
+    prompt: true
+  - name: platform
+    type: string
+    description: "Target: web, mobile, desktop, api, fullstack (default: fullstack)"
+    required: false
+    default: fullstack
+mcp_servers:
+  - fetch
+  - sequential-thinking
+  - memory
+  - ruflo
+---
+
+# /idea — From Thought to Blueprint in 60 Seconds
+
+Transform a single sentence into a complete product blueprint.
+
+## What Happens
+
+```
+"I want a Slack bot that summarizes PRs daily"
+                    │
+                    ▼
+┌─────────────────────────────────────────────────┐
+│  1. RESEARCH (fetch + ruflo swarm)               │
+│     • Search existing solutions                  │
+│     • Identify gaps and differentiators          │
+│     • Analyze target user personas               │
+│     • Estimate market size                       │
+├─────────────────────────────────────────────────┤
+│  2. REQUIREMENTS (sequential-thinking)           │
+│     • User stories with acceptance criteria      │
+│     • Functional requirements                    │
+│     • Non-functional requirements                │
+│     • Edge cases and constraints                 │
+├─────────────────────────────────────────────────┤
+│  3. ARCHITECTURE                                 │
+│     • System design (mermaid diagrams)           │
+│     • Data model (entities, relationships)       │
+│     • API design (endpoints, schemas)            │
+│     • Infrastructure (Cloudflare/Vercel/AWS)     │
+├─────────────────────────────────────────────────┤
+│  4. DESIGN SYSTEM                                │
+│     • Color palette, typography, spacing         │
+│     • Component inventory                        │
+│     • Page wireframes (text-based)               │
+│     • Apple HIG compliance notes                 │
+├─────────────────────────────────────────────────┤
+│  5. IMPLEMENTATION PLAN                          │
+│     • Sprint breakdown (2-week sprints)          │
+│     • Task dependency graph                      │
+│     • Tech stack recommendations                 │
+│     • Estimated complexity per task              │
+├─────────────────────────────────────────────────┤
+│  6. SCAFFOLD                                     │
+│     • Generate project structure                 │
+│     • Create package.json / pyproject.toml       │
+│     • Set up CI/CD templates                     │
+│     • Initialize git repo                        │
+└─────────────────────────────────────────────────┘
+```
+
+## Usage
+
+```bash
+/idea "AI-powered invoice scanner for freelancers"
+/idea "Real-time multiplayer code editor" --platform web
+/idea "Fitness tracker with Apple Watch sync" --platform mobile
+/idea "CLI tool that converts Figma to React components" --platform desktop
+```
+
+## In Pipes
+
+```bash
+/pipe idea "my idea" >> plan >> go *10 >> test >> ship
+/pipe idea "my idea" >> des >> hig >> brand >> go *5 >> browser-test >> ship
+```
+
+## Output
+
+```
+.luna/{project}/
+  idea/
+    research.md           # Market analysis, competitors, gaps
+    requirements.md       # User stories, acceptance criteria
+    architecture.md       # System design, data model, APIs
+    design-system.md      # Colors, typography, components
+    implementation-plan.md # Sprints, tasks, dependencies
+```

package/commands/ll-imagine.md

@@ -0,0 +1,72 @@
+---
+name: ll-imagine
+displayName: Luna Imagine
+description: AI image and visual generation — product screenshots, marketing assets, social cards, diagrams, 3D models, brand assets
+version: 1.0.0
+category: creative
+agent: luna-task-executor
+parameters:
+  - name: what
+    type: string
+    description: What to generate — description, or preset like "hero", "social-card", "og-image", "3d-model", "icon-set"
+    required: true
+    prompt: true
+  - name: provider
+    type: string
+    description: "Provider: auto (best for task), dalle, stable-diffusion, flux, midjourney, replicate"
+    required: false
+    default: auto
+  - name: style
+    type: string
+    description: "Style: photorealistic, illustration, 3d, minimal, abstract, branded"
+    required: false
+    default: branded
+mcp_servers:
+  - stability-ai
+  - piapi
+  - replicate
+  - fal-ai
+  - tripo-3d
+  - zai-mcp-server
+  - memory
+---
+
+# /imagine — See It Before You Build It
+
+Generate any visual asset your product needs — from hero images to 3D models to social cards. Uses the best AI model for each task.
+
+## Presets
+
+```bash
+/imagine hero "AI agents working in harmony"              # Hero section background
+/imagine social-card "LunaOS v2.0 Launch"                # OG image for social
+/imagine og-image                                          # Auto from product metadata
+/imagine icon-set "workflow, agent, deploy, monitor"      # App icon set (SVG)
+/imagine 3d-model "futuristic dashboard terminal"          # 3D asset (Tripo/Meshy)
+/imagine mockup "phone showing our app"                   # Device mockup
+/imagine diagram "our microservices architecture"          # Visual architecture diagram
+/imagine avatar "friendly AI assistant"                    # Brand mascot
+/imagine pattern "geometric luna theme"                    # Background pattern
+/imagine logo-variations                                   # Logo in different contexts
+```
+
+## Provider Auto-Selection
+
+| Task | Best Provider | Why |
+|------|--------------|-----|
+| Photorealistic images | Flux (via fal.ai) | Best photorealism |
+| Illustrations | Midjourney (via PiAPI) | Best artistic quality |
+| Quick iterations | DALL-E | Fastest generation |
+| Product mockups | Stable Diffusion | Best control with ControlNet |
+| 3D models | Tripo AI | Best text-to-3D |
+| Social cards | Flux + typography | Brand-consistent |
+
+## In Pipes
+
+```bash
+/pipe imagine hero >> brand >> hig >> browser-test        # Generate then verify
+/pipe imagine social-card >> publish twitter               # Social card → post
+/pipe imagine 3d-model "product" >> present pitch          # 3D asset for pitch deck
+/pipe ghost blog "topic" >> imagine hero >> publish medium  # Blog with hero image
+/pipe brand >> imagine icon-set >> imagine avatar >> ship   # Full brand asset kit
+```

package/commands/ll-inbox.md

@@ -0,0 +1,116 @@
+---
+name: ll-inbox
+displayName: Luna Inbox Zero
+description: AI email management — auto-categorize, bulk unsubscribe, draft replies, meeting briefs, and inbox analytics
+version: 1.0.0
+category: productivity
+agent: luna-task-executor
+parameters:
+  - name: action
+    type: string
+    description: "Action: clean (bulk unsubscribe), draft (auto-reply drafts), brief (meeting prep), analyze (inbox stats), triage (categorize)"
+    required: true
+    prompt: true
+  - name: account
+    type: string
+    description: Email account (Gmail, Outlook)
+    required: false
+    default: default
+prerequisites:
+  - Gmail or Outlook OAuth configured
+---
+
+# Luna Inbox Zero — AI Email Management
+
+Manage your inbox with AI — categorize emails, bulk unsubscribe, draft replies, prepare meeting briefs, and get inbox analytics.
+
+## What This Command Does
+
+1. **Triage** — Categorize unread emails: action-required, FYI, newsletter, spam, meeting
+2. **Clean** — Bulk unsubscribe from newsletters and marketing in one command
+3. **Draft** — AI pre-drafts replies for action-required emails
+4. **Brief** — Prepare meeting briefs from calendar + related email threads
+5. **Analyze** — Inbox analytics: volume, response time, top senders, busiest hours
+
+## Usage
+
+```bash
+# Categorize inbox
+/inbox triage
+
+# Bulk unsubscribe from newsletters
+/inbox clean
+
+# Draft replies for important emails
+/inbox draft
+
+# Prepare briefs for today's meetings
+/inbox brief
+
+# Inbox analytics
+/inbox analyze
+```
+
+## How It Works
+
+### Triage
+- Connects to Gmail/Outlook via OAuth
+- Reads unread emails (subject + first 200 chars)
+- Sends to Claw Gateway AI for categorization
+- Labels emails: action-required, FYI, newsletter, meeting, spam
+- Outputs prioritized list
+
+### Clean
+- Identifies all newsletter/marketing senders
+- Shows list with unsubscribe counts
+- Bulk unsubscribes (one-click via List-Unsubscribe header)
+- Archives old newsletters
+
+### Draft
+- Takes action-required emails
+- AI generates draft replies based on context
+- Saves as drafts (doesn't send)
+- You review and send
+
+### Brief
+- Reads today's calendar events
+- Finds related email threads per meeting
+- AI generates a 3-bullet brief per meeting
+- Includes: attendees, context, suggested talking points
+
+### Analyze
+- Inbox volume over time
+- Average response time
+- Top 10 senders
+- Busiest hours/days
+- Unread count trend
+
+## Output
+
+```
+.luna/{project}/inbox/
+  triage.md         # Categorized email list
+  unsubscribe.md    # Newsletter cleanup report
+  drafts/           # AI-drafted replies
+  briefs/           # Meeting preparation briefs
+  analytics.md      # Inbox stats
+```
+
+## Privacy
+
+- Emails processed via Claw Gateway AI (your own infrastructure)
+- No data sent to third parties
+- Self-hostable with Inbox Zero open-source (github.com/elie222/inbox-zero)
+- OAuth tokens encrypted at rest
+
+## In Pipes
+
+```bash
+# Morning routine: triage, brief, draft
+/pipe inbox triage >> inbox brief >> inbox draft
+
+# Weekly cleanup
+/pipe inbox clean >> inbox analyze
+```
+
+Reference: [Inbox Zero](https://github.com/elie222/inbox-zero)

package/commands/ll-lam.md

@@ -0,0 +1,93 @@
+---
+name: ll-lam
+displayName: Luna LAM (Large Action Model)
+description: Action-driven AI — combine RAG understanding with autonomous code actions across the full stack
+version: 1.0.0
+category: ai
+agent: luna-rag-enhanced
+parameters:
+  - name: goal
+    type: string
+    description: High-level goal in natural language
+    required: true
+    prompt: true
+  - name: scope
+    type: string
+    description: Project or feature scope
+    required: true
+    prompt: true
+workflow:
+  - understand_goal_with_rag
+  - plan_action_sequence
+  - execute_actions_autonomously
+  - validate_each_action
+  - loop_until_goal_met
+  - generate_lam_report
+output:
+  - .luna/{current-project}/lam-report.md
+prerequisites: []
+---
+
+# Luna LAM (Large Action Model)
+
+Goal-driven AI that understands your codebase (RAG) and takes autonomous actions to achieve objectives.
+
+## What This Command Does
+
+Combines semantic code understanding (Nexa RAG) with autonomous action execution to accomplish high-level goals. Unlike `/feature` which follows a fixed pipeline, LAM dynamically decides what actions to take based on codebase analysis.
+
+1. **Understand** — uses RAG to deeply understand your codebase and the goal
+2. **Plan** — generates an action sequence (not just code — files, configs, tests, deps)
+3. **Execute** — takes actions autonomously (create, edit, delete, install, configure)
+4. **Validate** — checks each action against rules and tests
+5. **Adapt** — if something fails, re-plans and tries alternative approach
+6. **Loop** — continues until the goal is met or reports blockers
+
+## How LAM Differs from Other Commands
+
+| Command | Approach |
+|---------|----------|
+| `/feature` | Fixed pipeline: plan → implement → test → review |
+| `/go` | Executes one task from a pre-made plan |
+| `/openhands` | Delegates to external autonomous agent |
+| `/lam` | **Dynamic actions** — decides what to do based on RAG understanding |
+
+## Usage
+
+```
+/lam "make the dashboard load 2x faster"
+/lam "add multi-tenancy support to the entire API"
+/lam "convert all API endpoints from REST to tRPC"
+/lam "set up a complete CI/CD pipeline with staging and production"
+```
+
+## Action Types LAM Can Take
+
+- Create files (components, services, routes, tests)
+- Edit existing files (refactor, optimize, add features)
+- Install dependencies (`npm install`, `pip install`)
+- Configure tools (ESLint, Tailwind, Prisma, Docker)
+- Run commands (build, test, migrate, deploy)
+- Generate documentation
+- Create database migrations
+- Set up infrastructure configs
+
+## Architecture
+
+```
+Natural language goal
+  -> RAG: understand codebase context
+  -> Planner: generate action sequence
+  -> Executor: take actions
+  -> Validator: check results
+  -> Feedback loop: adapt if needed
+  -> Goal achieved
+```
+
+## Safety
+
+- Validates every action against project rules (100-line cap, tests)
+- Dry-run mode available for destructive actions
+- Checkpoints after each major action for rollback
+- Maximum 10 action iterations before reporting blockers
+- All changes are reviewable before commit

package/commands/ll-landing.md

@@ -0,0 +1,171 @@
+---
+name: ll-landing
+displayName: Luna Landing Page
+description: Generate a HeyGen-quality marketing landing page for any product — rotating 3D hero, social proof counters, logo carousel, feature tabs, testimonials, FAQ, mega footer
+version: 1.0.0
+category: creative
+agent: luna-task-executor
+parameters:
+  - name: product
+    type: string
+    description: Product name
+    required: true
+    prompt: true
+  - name: description
+    type: string
+    description: One-line product description
+    required: true
+    prompt: true
+  - name: style
+    type: string
+    description: "Style: heygen (dark hero + white body), dark (full dark), light (full light), minimal"
+    required: false
+    default: heygen
+  - name: accent
+    type: string
+    description: "Brand accent color hex (e.g., #00C8FF)"
+    required: false
+    default: "#7c5cfc"
+prerequisites: []
+---
+
+# Luna Landing Page — HeyGen-Quality Marketing Pages
+
+Generate a professional SaaS marketing landing page matching the quality of HeyGen, Linear, Vercel — with rotating 3D hero, animated counters, logo carousel, tabbed features, testimonials, and mega footer.
+
+## What This Command Does
+
+1. **Analyzes** your product (reads codebase, README, or URL)
+2. **Generates** a complete Next.js + Tailwind CSS landing page with 18 sections
+3. **Creates** a rotating 3D hero (CSS cube or video embed)
+4. **Writes** conversion-optimized copy using Claw Gateway AI
+5. **Outputs** production-ready code, deployable to Cloudflare Pages
+
+## Page Sections (in order)
+
+### 1. Sticky Nav Bar
+Logo left, nav links center, Sign In + CTA right. Clean, minimal height.
+
+### 2. Hero Section (split layout)
+- **Left**: Large bold headline in accent gradient, 2-3 sentence sub, Google SSO button + primary CTA
+- **Right**: Rotating 3D diamond/cube `<video>` element (or CSS `preserve-3d` fallback)
+- White/dark background, generous padding, vertically centered
+
+### 3. Social Proof Counter Bar
+Dark rounded card with 3 animated count-up numbers on scroll (e.g., "10M+ users", "50K+ projects"). Uses `IntersectionObserver` for trigger.
+
+### 4. Logo Carousel (Trust Bar)
+"Trusted by" label + infinite auto-scrolling grayscale logos. CSS `translateX` animation, pause on hover. Duplicated for seamless loop.
+
+### 5. Product Feature Cards
+Large heading + horizontal scroll or grid of cards with thumbnail, tag, title, description. Hover lift effect.
+
+### 6. Tabbed Feature Showcase
+Row of tab buttons with icons. Click swaps: icon, title, heading, description, CTA. `fadeSlideIn` animation.
+
+### 7. Comparison Section
+Side-by-side "Us vs Them" with visual proof. Sub-features with heading + description + link.
+
+### 8. Demo/Video Section
+Heading + description + embedded video player. Language/variant switch buttons. Feature highlights below.
+
+### 9. Studio/Tool Section
+Banner heading + CTA. Accordion or tabs showing key features.
+
+### 10. Customer Testimonial
+Blockquote + name + title + company. Large stat callout. Photo.
+
+### 11. Second Logo Carousel
+
+### 12. Vision/Mission Section
+Short mission statement + "Learn more" CTA + decorative backgrounds.
+
+### 13. Use Case Cards Grid
+4 cards: different audiences/industries. Each with heading, icon, image, description.
+
+### 14. API/Integration Banner
+Floating tech icons + heading + "Explore" CTA.
+
+### 15. Compliance Badges
+Row of certification logos (SOC 2, GDPR, etc.) with labels.
+
+### 16. FAQ Accordion
+15+ collapsible Q&A items. Generated from product context.
+
+### 17. Final CTA Section
+"Start [action] today" heading + CTA button + gradient backgrounds.
+
+### 18. Mega Footer
+Multi-column: logo, language selector, Products, Resources, Company, Legal. Social icons. Copyright.
+
+## CSS Rotating Diamond Hero
+
+```css
+.cube-scene { perspective: 800px; }
+.cube {
+  transform-style: preserve-3d;
+  animation: rotateCube 8s ease-in-out infinite;
+  transform: rotateX(-20deg) rotateZ(45deg);
+}
+.cube__face {
+  border-radius: 16px;
+  backface-visibility: hidden;
+  box-shadow: 0 0 20px rgba(0,255,200,0.2);
+  border: 1.5px solid rgba(0,255,200,0.25);
+}
+@keyframes rotateCube {
+  0%   { transform: rotateX(-20deg) rotateZ(45deg) rotateY(0deg); }
+  25%  { transform: rotateX(-20deg) rotateZ(45deg) rotateY(90deg); }
+  50%  { transform: rotateX(-20deg) rotateZ(45deg) rotateY(180deg); }
+  75%  { transform: rotateX(-20deg) rotateZ(45deg) rotateY(270deg); }
+  100% { transform: rotateX(-20deg) rotateZ(45deg) rotateY(360deg); }
+}
+```
+
+## Key Design Patterns
+
+1. **Color**: White/light body, dark accents, single strong accent color for headings + CTAs
+2. **Typography**: Large bold display font for hero, clean sans-serif for body, big numbers for proof
+3. **Animations**: Count-up on scroll, infinite logo carousel, tab fade-slide, SVG stroke draw
+4. **Hero video**: Pre-rendered 3D loop as `<video autoplay muted loop playsinline>` (best quality)
+5. **Layout**: Two-column hero flexbox, alternating full-width and contained sections
+
+## Usage
+
+```bash
+# Generate for your product
+/landing "LunaOS" "AI-Native Backend Platform for Developers"
+
+# With custom style
+/landing "CodeRailFlow" "Browser automation platform" --style dark --accent "#3b82f6"
+
+# Minimal landing page
+/landing "MyApp" "Project management for teams" --style minimal
+```
+
+## Output
+
+```
+.luna/{project}/landing/
+  index.html          # Complete landing page
+  styles.css          # Tailwind + custom CSS
+  components/         # React components (if Next.js)
+  assets/
+    hero-cube.mp4     # 3D hero video (or CSS fallback)
+    logos/             # Trust logos (placeholder)
+  copy.json           # All generated copy
+  deploy.sh           # Cloudflare Pages deploy script
+```
+
+## In Pipes
+
+```bash
+# Generate landing + deploy
+/pipe landing "MyApp" "description" >> ship
+
+# Generate landing + record demo + produce video
+/pipe landing "MyApp" "desc" >> flow-record http://localhost:3000 >> heygen http://localhost:3000
+
+# Full launch: landing + audit + fix + deploy
+/pipe landing "MyApp" "desc" >> site-audit http://localhost:3000 >> fix >> ship
+```