AGENT.md

Build/Test Commands

• bundle exec jekyll serve - Start development server (http://localhost:4000)
• bundle exec jekyll build - Build static site to _site/
• bundle install - Install/update dependencies
• No specific test framework - this is a static Jekyll blog

Architecture & Structure

• Jekyll static site using minimal-mistakes theme with GitHub Pages deployment
• Content types: Blog posts (_posts/), projects (_projects/), linklog (_linklog/), pages (_pages/)
• Collections: Posts, projects, linklog all have permalinks and feed generation
• Plugins: jekyll-paginate, jekyll-sitemap, jekyll-gist, jekyll-feed, jemoji, jekyll-include-cache
• Analytics: Google Analytics (G-KQGNGZL7KQ)

Code Style & Content Guidelines

• Front matter: YAML with title, categories (Blog/Project), tags, optional image/excerpt
• File naming: YYYY-MM-DD-title-with-hyphens.md for posts
• Images: Store in /assets/images/ and reference with full paths

Content Writing Style (from STYLE_GUIDE.md)

• Technical precision with accessibility: Complex security concepts made understandable without oversimplification
• Memorable metaphors: Use vivid analogies (“remediation ballet”, airplane cockpit for cloud configs)
• Informal authority: Self-deprecating humor (“two blog readers”, “unfinished thoughts”) while maintaining credibility
• BLUF approach: Start with key findings, follow with context, end with actionable recommendations
• Problem-solution framework: Identify challenge → technical analysis → practical examples → mitigation strategies
• Focus areas: AI/GenAI security, low-code/no-code security, Microsoft 365 ecosystem, “democratization risks”
• Key terminology: “promptware”, “remote copilot execution”, “citizen developers”, “living-off-the-land”
• Characteristic phrases: “without paying for it in security incidents”, “the elephant in the room”
• Tone: Professional yet approachable, collaborative (“we”), attacker perspective, emphasize “what to do about it”