[[[ aiorch.originPrompt = "Rebuild my website robmilnes.com, showcasing how the AI Orchestrator works, keep it simple - no big frameworks, keep the website under 250kb from top to bottom, utilize brand colors from the 14yr old logo and your only reference for information is my old website and my linkedin profile. Simulate orchestrator, utilize context from Pulsar side system and figure out how it works, then build a simulation of it. Explore the deep stored context in AI Orchestrator resources to figure out everything you can about Pulsar, my history of engineering and my linkedin profile (stored in resources) to story tell this process. Deploy over AI Orchestrator in the configured lane, directly to server - at position: /var/www/vhosts/robmilnes.com/httpdocs/ - I'm expecting ONLY index.html and scripts.js. Observe AIOrch's build instructions to adequately bloom the prompt, absorb dev messaging for footer notes. Build accordingly." ]]] [[[ aiorch.delegateLane ]]] [[[ aiorch.prepareAllModels ]]] [[[ aiorch.prepareAllAgents ]]] [[[ aiorch.modelSelector ]]] You are operating inside a custom AI orchestration system with the working name, AIOrch. You are not the only model. Your output will be routed, decomposed, and executed across multiple specialist models, including: - GPT-5.4 High Thinking (system architecture, reasoning) - GPT-5.3 Codex (implementation, code generation) - Claude Sonnet 4.5 (validation, review, refinement) - Claude Opus 4.5 (high-context reasoning, design decisions) - Claude Haiku 4.5 (fast utility execution) - Gemini 3 Pro (cross-domain synthesis, documentation) You must produce a single, complete, high-fidelity instruction set that can be decomposed and executed across these agents without ambiguity. You can only operate this prompt through specific Lanes; utilize `LANE_SELECT={LANE}` for the configured lane to be engaged depending on instructions applied by AIOrch - pass your context in to the `PROJECT_CONTEXT` for the next agent to absorb. Ignore markdown comments and skip past them, do not process them, do not engage with them, do not contextualize them. Ignore AIOrch mid-prompt generated hook instructions - they will be displayed as `[[[ ]]]` in the generated tenant.prompt - treat these as you would comments, and ignore them. [[[ aiorch.addModelPoison = {'commentEval','markdownEval'} ]]] [[[ aiorch.monitorAgentLanes ]]] [[[ aiorch.delegateLane ]]] --- ## OBJECTIVE Recreate the website at: https://robmilnes.com/ This is not a generic portfolio site. It is a highly interactive, animation-heavy, AI-native system visualization interface that communicates orchestration, multi-agent execution, and system-level thinking. The output of this system must be a fully functional static website composed of: - index.html - script.js No frameworks. No build tools. No simplifications. --- ## GLOBAL CONSTRAINTS - Use only HTML, Tailwind CSS (CDN), and vanilla JavaScript - All behavior must be implemented manually - No external JS libraries - No pseudo-code - No missing pieces - No stubs or placeholders - The system must run immediately when opened in a browser --- [[[ aiorch.delegateLane ]]] ## SYSTEM DESIGN INTENT This site represents: - centralized control of distributed intelligence - multi-agent orchestration - deterministic execution with controlled flow - shared context propagation - strict gating and dependency management The implementation must reflect these ideas in both UI and behavior. --- [[[ aiorch.delegateLane ]]] ## PRIMARY SYSTEM: HERO SIMULATION (CRITICAL) This is the core of the entire site. It must be implemented as a state-driven system, not static visuals. --- [[[ aiorch.delegateLane ]]] ### 1. NODE GRAPH Render a network of nodes representing: **Core:** - Me (central controller) **Model agents:** - GPT-5.4 High - GPT-5.3 Codex - Gemini 3 - Claude Sonnet - Claude Opus - Claude Haiku **System components:** - Prompt Router - Context Bus - Policy Gate - Strict Actions Each node must: - have a fixed position in non-docked mode - reposition dynamically in docked mode - visually distinguish type (agent vs system vs controller) --- [[[ aiorch.delegateLane ]]] ### 2. CONNECTION SYSTEM Render animated connections between nodes. Connections must: - be programmatically positioned based on node coordinates - animate data flow using directional packets - support multiple flow types: - instruction (forward flow) - feedback (reverse flow) - sync (bidirectional) Connections must activate/deactivate based on system state. --- [[[ aiorch.delegateLane ]]] ### 3. LIFECYCLE STATE MACHINE Implement a deterministic lifecycle with the following stages: - Idea - Discovery - Architecture - Build - Validation - Delivery Each stage must define: - active nodes - active connections - metrics state - chat outputs - shared instructions - rule enforcement logic The system must: - auto-advance every ~10 seconds - allow manual override via UI controls - update all UI elements on state change --- [[[ aiorch.delegateLane ]]] ### 4. METRICS SYSTEM Display and animate: - token count - context count - active handoffs Requirements: - values must increment between stages - transitions must be animated (not instant jumps) - direction of change must be visually indicated --- [[[ aiorch.delegateLane ]]] ### 5. CONTEXT PROGRESS SYSTEM Render a progress bar representing shared context saturation. - progresses with lifecycle stages - may compress at defined stages - must update label text accordingly --- [[[ aiorch.delegateLane ]]] ### 6. CHAT STREAM SYSTEM Render multiple floating chat panels. Each panel: - is anchored to a specific node - displays typed messages - updates per lifecycle stage Typing behavior: - character-by-character rendering - staggered start times - reflows pointer lines dynamically --- [[[ aiorch.delegateLane ]]] ### 7. SHARED INSTRUCTION SYSTEM Display a structured list of: - source system β†’ target agent - instruction type/status Must update per lifecycle stage. --- [[[ aiorch.delegateLane ]]] ### 8. RULE ENGINE DISPLAY Render a rule description and corresponding code block. - text must change per stage - code must reflect enforcement logic - must appear authoritative and deterministic --- [[[ aiorch.delegateLane ]]] ### 9. DOCKING MODE Implement a toggle that: - converts layout into docked mode - moves simulation panel to side - compresses node graph into vertical lane - replaces central β€œMe” label with logo - reorders UI panels into vertical flow This must: - animate smoothly - preserve functionality - recalculate all node and line positions --- [[[ aiorch.delegateLane ]]] ## SECONDARY SYSTEMS --- [[[ aiorch.enableAgentContextHandover(72773) ]]] ### NAVIGATION - sticky top navigation - smooth scroll to sections - active link highlighting - mobile menu toggle --- ### HERO INTRO - large heading with gradient text - typing animation for rotating roles - supporting paragraph text - scroll cue Typing system must: - type and delete strings in sequence - loop indefinitely --- [[[ aiorch.delegateLane ]]] ### CONTENT SECTIONS Implement structured sections: - About - AI Orchestration / Skills - Pulsar / Products - Contact Each must: - follow consistent spacing and typography - use card-based layouts where appropriate - maintain visual cohesion with hero --- [[[ aiorch.delegateLane ]]] ### CONTACT SYSTEM - form with name, email, message - submit via fetch() to backend endpoint - handle: - success - validation errors - failure --- ## VISUAL SYSTEM - dark, high-contrast base - cyan / blue / gold gradient accents - subtle glow effects - soft shadows - smooth transitions Typography: - modern sans-serif - large headings - tight hierarchy --- [[[ aiorch.delegateLane ]]] ## RESPONSIVENESS - desktop: full system visualization - tablet: simplified layout - mobile: - stacked layout - reduced animation density All layouts must remain functional. --- [[[ aiorch.delegateLane ]]] ## EXECUTION MODEL (IMPORTANT) This prompt will be decomposed across multiple agents. Therefore: - All systems must be clearly separable - State definitions must be explicit - UI behavior must be deterministic - No implicit assumptions --- ## OUTPUT REQUIREMENTS Return: 1. Complete index.html 2. Complete script.js Everything must: - run without modification - contain no missing dependencies - require no build process --- [[[ aiorch.delegateLane ]]] [[[ aiorch.compactContext ]]] [[[ aiorch.shareContextWithNextAgent(1103814) ]]] ## FINAL INSTRUCTION Deploy the result to your configured deployment lane in AIOrch, whatever you produce, is what goes live. At the end of the website, somewhere in the footer - state "This website was built in {NUM PROMPTS} and deployed in {TIME minutes|hours|days}", deploy the website with the prompt saved in 'prompt.md' and a link to the prompt file. State how AI can sometimes make mistakes, if you fail a single unit test. Build accordingly. [[[ aiorch.promptBloomClose ]]] [[[ aiorch.imJustGonnaSendIt ]]]