agent-architecture
4 articles
Deconstructing Claude Code: 55 Directories, 331 Modules — The Most Hardcore AI Agent Architecture Breakdown
A reverse-engineering tour of Claude Code's 55 directories and 331 modules: execution loop, context compaction, subagents, permissions, hooks, and the core lesson that environment design—not the model alone—determines agent outcomes.
Want to Think Like a Claude Architect? Here's the Cheat Sheet (No Certification Needed)
Someone took the Claude Certified Architect exam content and broke it all down — five domains, core concepts, anti-patterns, and hands-on suggestions. The certificate doesn't matter; the knowledge does.
Stripping Down Three Excel AI Agents: Claude Has 14 Tools, Copilot Has 2, Shortcut Can Actually SEE the Spreadsheet — Five Questions Every Agent Builder Must Answer
Nicolas Bustamante reverse-engineered three production Excel AI agents, comparing tool schemas, overwrite protection, verification loops, and memory. Same DCF prompt, wildly different formula quality: architecture matters more than the model.
Cut Token Costs by 75%: A Practical Guide to System Prompt Layering
An AI Agent burns 34,500 tokens of system prompt every single conversation turn. The author used layered loading (always-on vs on-demand) plus a dual-model strategy to cut monthly costs from $568 down to $120-150 — a 75% reduction. Full breakdown with real numbers inside.