@leanspec/ui 0.2.5 → 0.2.6-dev.20251125010539

package/.next/standalone/packages/ui/dist/standalone/specs/043-official-launch-02/DEMO-AI-ASSISTED.md

@@ -1,359 +0,0 @@
-# AI-Assisted Demo Script
-
-## Overview
-
-Showcase LeanSpec's **killer feature**: AI agents working with specs naturally through MCP integration.
-
-**Core Message:** *"Specs that AI agents can read, write, and reason about—automatically."*
-
-**Duration:** 3-5 minutes  
-**Format:** Split-screen (AI chat + terminal)  
-**Recording:** OBS Studio / Loom with split view
-
----
-
-## Screen Layout
-
-```
-┌─────────────────────────────────────────────────────────┐
-│  AI Chat (Claude / Copilot)     │  Terminal + Editor   │
-│         60% width                │      40% width        │
-│                                  │                       │
-│  Show natural conversation       │  Show real-time      │
-│  with MCP tool calls visible     │  file changes        │
-└─────────────────────────────────────────────────────────┘
-```
-
----
-
-## Full Script
-
-### SCENE 1: Setup & Discovery (45 seconds)
-
-**[Terminal - Right side]**
-```bash
-# Show clean terminal
-cd ~/demo-project
-npm install --save-dev lean-spec
-npx lean-spec init  # Choose "minimal"
-```
-
-**[AI Chat - Left side]**
-```
-Human: "What specs exist in this project?"
-
-AI: [Uses MCP tool: list]
-    "The project currently has no specs. The specs/ directory 
-     is initialized and ready. Would you like to create one?"
-```
-
-**Key Point:** AI has instant awareness of project state through MCP
-
----
-
-### SCENE 2: Create Spec Through Conversation (90 seconds)
-
-**[AI Chat]**
-```
-Human: "We need user authentication. Create a spec for JWT-based auth 
-        with bcrypt password hashing and rate limiting."
-
-AI: [Uses MCP: create("user-authentication")]
-    "I've created spec 001-user-authentication. 
-     Let me draft the content following LeanSpec principles..."
-    
-    [Uses MCP: view("user-authentication")]
-    [AI edits the spec file directly]
-    
-    "I've added:
-    
-    **Problem**: No authentication, unprotected endpoints
-    
-    **Solution**: 
-    - JWT tokens (24hr expiry)
-    - bcrypt hashing (10 rounds minimum)
-    - Rate limiting (5 attempts/min per IP)
-    
-    **Success Criteria**:
-    - [ ] Users can register with email/password
-    - [ ] Login returns valid JWT token
-    - [ ] Protected endpoints require valid token
-    - [ ] Rate limit blocks brute force attacks
-    
-    The spec is under 200 lines and ready for review."
-```
-
-**[Terminal/Editor - Right side]**
-- Show `specs/001-user-authentication/README.md` appearing
-- Show content being written in real-time
-- Highlight the clean frontmatter structure
-
-**Key Point:** AI drafts complete, actionable spec in seconds
-
----
-
-### SCENE 3: Semantic Search & Memory (60 seconds)
-
-**[AI Chat]**
-```
-Human: "What did we decide about password security?"
-
-AI: [Uses MCP: search("password security")]
-    [Uses MCP: view("001-user-authentication")]
-    
-    "Per spec 001-user-authentication:
-    - Passwords must be hashed with bcrypt
-    - Minimum 10 rounds (computationally expensive to crack)
-    - Never store plaintext passwords
-    - Rate limiting prevents brute force (5 attempts/min)"
-```
-
-**Key Point:** AI has semantic memory—no need to reference file paths
-
----
-
-### SCENE 4: Update Status & Track Progress (45 seconds)
-
-**[AI Chat]**
-```
-Human: "I'm starting work on the auth spec. Update its status."
-
-AI: [Uses MCP: update("user-authentication", status="in-progress")]
-    "Spec 001 status updated to 'in-progress'.
-     Transition logged at 2025-11-10T18:30:00Z"
-    
-Human: "Show me the project board"
-
-AI: [Uses MCP: board()]
-    "Project Board:
-    
-    📅 Planned (0)
-    ⏳ In Progress (1)
-      - 001-user-authentication [in-progress]
-    ✅ Complete (0)
-    
-    Project Completion: 0% (0/1 complete)"
-```
-
-**[Terminal - Right side]**
-```bash
-# Verify in terminal
-npx lean-spec board
-# Shows same Kanban view
-```
-
-**Key Point:** AI tracks project state automatically
-
----
-
-### SCENE 5: Validation & Quality Gates (45 seconds)
-
-**[AI Chat]**
-```
-Human: "I added a lot of content to the spec. Is it getting too complex?"
-
-AI: [Uses MCP: validate()]
-    "Validation Results:
-    
-    ✅ 001-user-authentication: 187 lines (healthy)
-    
-    All specs pass validation. Your spec is well-scoped and 
-    under the 300-line warning threshold. No issues detected."
-```
-
-**[Simulate adding too much content]**
-
-```
-Human: "What if I add 300 more lines of implementation details?"
-
-AI: [Uses MCP: validate()]
-    "⚠️ Validation Warnings:
-    
-    001-user-authentication: 487 lines (warning: exceeds 400)
-    
-    Recommendation: This spec exceeds the Context Economy threshold.
-    Consider splitting into sub-specs:
-    - README.md (overview)
-    - DESIGN.md (detailed architecture)
-    - IMPLEMENTATION.md (step-by-step plan)"
-```
-
-**Key Point:** AI enforces first principles automatically
-
----
-
-### SCENE 6: Wrap-Up (30 seconds)
-
-**[AI Chat]**
-```
-Human: "Show me all available commands"
-
-AI: [Uses MCP: help or lists tools]
-    "I can help you with these spec operations:
-    
-    **Discovery**: list, search, board, stats, deps
-    **Viewing**: view (formatted/raw/json), files
-    **Management**: create, update, archive
-    **Quality**: validate, check
-    
-    All through natural conversation—no need to remember CLI syntax."
-```
-
-**[Voiceover or text overlay]**
-```
-"LeanSpec + MCP: Specs that AI agents understand natively.
-
-- Natural conversation interface
-- Semantic memory across sessions  
-- Automatic validation and quality gates
-- Zero-friction collaboration
-
-Get started: npm install lean-spec"
-```
-
----
-
-## Key Visual Moments (For Editing)
-
-1. **0:15** - MCP tool call appears (highlight with subtle animation)
-2. **1:30** - Spec file appears in editor in real-time
-3. **2:15** - Search finding spec by concept, not filename
-4. **3:00** - Status update reflected in board instantly
-5. **3:45** - Validation warning catching complexity
-
----
-
-## Voiceover Script (Complete)
-
-```
-[0:00 - INTRO]
-"LeanSpec is a spec methodology designed for AI-powered development.
-But here's what makes it different: specs that AI agents can read, 
-write, and manage—automatically.
-
-[0:15 - SETUP]
-Let me show you. I'll start by installing LeanSpec and asking 
-my AI assistant what specs exist in the project.
-
-Notice: The AI uses the Model Context Protocol to query LeanSpec 
-directly. No manual file references needed.
-
-[0:45 - CREATE]
-Now I'll ask the AI to create a spec for user authentication.
-Watch what happens.
-
-The AI creates the spec, understands the requirements, and drafts 
-complete content—all in natural language. Problem, solution, success 
-criteria. Clean, actionable, under 200 lines.
-
-[2:00 - SEMANTIC SEARCH]
-Days later, I can ask: 'What did we decide about password security?'
-
-The AI searches specs semantically, retrieves the authentication spec,
-and gives me the exact details. This works across sessions—
-the AI remembers project decisions through specs.
-
-[2:45 - STATUS TRACKING]
-When I start work, I just say: 'Update the auth spec to in-progress.'
-
-The AI updates the status and can show me the project board instantly.
-All without leaving the conversation.
-
-[3:30 - VALIDATION]
-Here's where it gets smart. If I add too much content, 
-the AI warns me—using LeanSpec's validation rules.
-
-It enforces the first principle: Context Economy. 
-Specs must fit in working memory, or they're split automatically.
-
-[4:15 - WRAP UP]
-This is LeanSpec with MCP: specs as semantic memory for AI agents.
-
-Natural conversation. Zero manual file management. Built-in quality gates.
-Perfect for human-AI collaboration.
-
-Try it yourself: npm install lean-spec"
-```
-
----
-
-## Recording Checklist
-
-### Pre-Recording
-- [ ] Clean terminal (no history, simple prompt)
-- [ ] Fresh project directory (`demo-project`)
-- [ ] AI tool configured (MCP server ready)
-- [ ] VS Code with simple theme (high contrast)
-- [ ] Screen recording software set to split view
-- [ ] Test MCP integration (verify tools work)
-- [ ] Prepare "overflow content" file for validation demo
-
-### During Recording
-- [ ] Slow down typing in chat (readable)
-- [ ] Pause 2-3 seconds after each AI response
-- [ ] Show MCP tool calls (if visible in UI)
-- [ ] Highlight key moments (spec file appearing, validation warnings)
-- [ ] Keep consistent pacing (not too fast)
-
-### Post-Production
-- [ ] Add subtle highlights for MCP tool calls
-- [ ] Zoom in on key text (validation warnings, spec content)
-- [ ] Add captions for key takeaways
-- [ ] Trim any dead air or mistakes
-- [ ] Add outro with installation command + website
-
----
-
-## Distribution
-
-### Primary
-- [ ] YouTube (unlisted for now, public at launch)
-- [ ] Embedded in README.md "Quick Demo" section
-- [ ] Docs site homepage (hero section)
-- [ ] Product Hunt submission (video)
-
-### Secondary
-- [ ] Twitter/X (15-30s teaser + link)
-- [ ] LinkedIn (professional audience)
-- [ ] Hacker News Show HN (link to full video)
-- [ ] Dev.to / Hashnode article embedding
-
----
-
-## Alternative: Recorded Terminal Session (Backup)
-
-If split-screen recording is complex, fall back to terminal-only demo with visible MCP output:
-
-```bash
-# Install
-npm install -g lean-spec
-lean-spec init
-
-# Show AI interaction via terminal (Claude CLI or similar)
-echo "Pretend this is AI chat, show the same workflow"
-
-# Manually run commands to show what AI would do via MCP
-lean-spec list
-lean-spec create user-authentication
-lean-spec view user-authentication
-lean-spec update user-authentication --status in-progress
-lean-spec board
-lean-spec validate
-```
-
-This loses the "natural conversation" magic but still shows the workflow.
-
----
-
-## Success Metrics
-
-**Demo is successful if viewers understand:**
-1. ✅ AI can manage specs through conversation (no CLI needed)
-2. ✅ Specs persist as semantic memory across sessions
-3. ✅ Validation enforces quality automatically
-4. ✅ MCP enables zero-friction AI collaboration
-
-**Failure signs:**
-- ❌ Looks like "just another CLI tool"
-- ❌ MCP integration not clear
-- ❌ Seems like gimmick, not real workflow

package/.next/standalone/packages/ui/dist/standalone/specs/043-official-launch-02/DEMO-FULL-LIFECYCLE.md

@@ -1,86 +0,0 @@
-# Full Lifecycle Demo: Spec → Implementation
-
-## Overview
-
-Show the **complete workflow** from AI-assisted spec creation through working implementation. This demonstrates LeanSpec's real-world value, not just tooling.
-
-**Duration:** 5-7 minutes  
-**Format:** Split-screen (AI chat + IDE with terminal)  
-**Outcome:** Working feature at the end
-
-**Core Message:** *"From conversation to code: How AI and specs work together"*
-
----
-
-## Demo Feature: Add `--json` Output to `lean-spec list`
-
-**Why this feature?**
-- ✅ Simple enough to implement in 5 minutes
-- ✅ Real feature users would want
-- ✅ Shows actual code changes
-- ✅ Has testable success criteria
-- ✅ Demonstrates validation workflow
-
-**Current state:** `lean-spec list` outputs formatted table  
-**Goal:** Add `--json` flag for machine-readable output
-
----
-
-## Full Script
-
-### SCENE 1: Project Context (30 seconds)
-
-**[Terminal]**
-```bash
-cd ~/lean-spec
-npx lean-spec list
-# Shows formatted table output
-```
-
-**[AI Chat]**
-```
-Human: "I need to add a --json flag to the 'list' command 
-        for machine-readable output. Help me create a spec."
-
-AI: "Let me check the existing specs first..."
-    [Uses MCP: search("list command")]
-    [Uses MCP: list()]
-    
-    "I don't see an existing spec for this. Let me create one 
-     following LeanSpec principles."
-```
-
----
-
-### SCENE 2: AI-Assisted Spec Creation (90 seconds)
-
-**[AI Chat]**
-```
-AI: [Uses MCP: create("json-output-flag")]
-    
-    "I've created spec 064-json-output-flag. Let me draft the content:
-    
-    **Problem**: 
-    - CLI output is human-readable only
-    - Can't pipe results to other tools
-    - No machine-readable format for automation
-    
-    **Solution**:
-    - Add --json flag to list command
-    - Output valid JSON array of specs
-    - Preserve existing formatting as default
-    
-    **Success Criteria**:
-    - [ ] lean-spec list --json outputs valid JSON
-    - [ ] JSON includes all spec metadata (name, status, tags, etc.)
-    - [ ] Existing behavior unchanged without flag
-    - [ ] Can pipe to jq successfully
-    
-    **Implementation**:
-    - Update src/commands/list.ts
-    - Add --json option to commander config
-    - Add JSON formatter function
-    - Add test cases
-    
-    Spec is 127 lines - well under Context Economy threshold.
-    Ready to implement?"