Nova Memory & Awareness Workflow

Nova’s awareness system is powered by a series of scripts that run daily or manually via GitHub Actions. These scripts analyze the codebase, extract meaningful signals, and generate awareness files.

• scanSiteStructure.js — Map folder + page structure
• componentScan.js — Index includes + layout modules
• imageAudit.js — Log hero/media/card assets
• codeFootprint.js — Count files + lines across codebase
• jsFunctionMap.js — Map function names in /scripts
• generateCodeMap.js — NEW: Code Awareness Sensor (JS/CSS/HTML map)
• unusedCSS.js — Identify potentially unused classes
• moodScan.js — Detect site tone and writing style
• aiPromptGenerator.js — Create poetic AI prompt of the day
• apiMonitor.js — Ping HuggingFace + Meme API
• changelogSnapshot.js — Pull last 10 Git commits
• versionTracker.js — Capture commit hash + timestamp
• generateNovaBoot.js — Combine all into Nova’s memory file

Each run generates awareness JSON files used in the dashboard and site logic:

• nova-session-boot.txt – Nova’s unified memory file (for GPT or local AI)
• nova-session-boot.html – Human-readable boot log
• code-map.json – NEW: Tag + function map across JS/CSS/HTML
• api-monitor.json, ai-prompts.json, mood-scan.json...

.github/workflows/update-nova-memory.yml
└── node scripts/*.js
    ├── analyze
    ├── write to /data/*.json
    └── push to Git → triggers Azure deploy
      

This system reads all JS, CSS, and HTML across the site and builds:

• Total counts of tags, class names, and functions
• Extracted function/class/tag names across all directories
• Timestamped entries for use in code-map.json

This output enables Nova to reflect on the evolving codebase, mention new functions, and generate targeted feedback or prompts.