🔬 Awesome AI Anatomy

11 AI agent projects dissected. Some had hidden pet systems. Some had zero security. All had God Objects.

Claude Code ships 18 virtual pet species in production. Pi Mono can impersonate Claude Code to dodge rate limits. MiroFish's "collective intelligence" is just LLMs role-playing on a fake social network. These aren't hot takes — they're findings from reading every line of code.

Most "awesome" lists link to repos. We crack them open — architecture diagrams, design trade-offs, code smells, security gaps, and the engineering decisions nobody talks about in the docs.

🔥 Greatest Hits

The 5 findings people can't stop sharing:

#	Finding	Project	Why It Matters
1	Claude Code ships 18 virtual pet species — a full tamagotchi system hidden in a coding agent	Claude Code	What else is hiding in tools you run with `sudo`?
2	Pi Mono's "stealth mode" impersonates Claude Code — fakes tool names to dodge rate limits	Pi Mono	Open-source projects gaming API providers is a trend
3	MiroFish's "collective intelligence" = zero collective intelligence — it's LLMs role-playing humans on a simulated social network	MiroFish	50K stars built on marketing, not on the tech the name promises
4	Lightpanda's bitcast dispatch makes Zig act like a language with vtables — a systems trick we'd never seen before	Lightpanda	Zig comptime metaprogramming pushed to its limits
5	DeerFlow's orphan tool call bug — 93 lines of code fix a crash that affects every LangGraph-based agent	DeerFlow	If you use LangGraph, you probably have this bug too

These are real findings from reading source code, not vibes from README files.

📊 Project Index

11 AI agent source code analyses — from orchestration frameworks to coding assistants.

#	Project	Stars	Language	Lines	Key Findings	Status
001	Dify	136K ⭐	Python + TS	1.24M	`graphon` engine, 7+ containers, 30+ vector DBs	✅ Published
002	Claude Code	109K ⭐	TypeScript	510K	4-layer context, hidden pet system (18 species)	✅ Published
003	DeerFlow	58K ⭐	Python + TS	—	14-layer middleware, hash loop kill@5, no RBAC	✅ Published
004	MiroFish	50K ⭐	Python + Vue	39K	LLM social sim as "collective intelligence"	✅ Published
005	Goose	37K ⭐	Rust + TS	198K	MCP-first bus, 5-inspector pipeline, 30+ LLMs	✅ Published
006	Pi Mono	32K ⭐	TypeScript	147K	Game-engine arch, stealth mode as Claude Code	✅ Published
007	Lightpanda	27K ⭐	Zig + Rust	91K	From-scratch browser for AI agents, 9× Chrome	✅ Published
008	Hermes Agent	26K ⭐	Python	260K	Self-improving skills, FTS5 search, frozen mem	✅ Published
009	oh-my-claudecode	24K ⭐	TypeScript	194K	19-agent team via file IPC, Haiku→Opus routing	✅ Published
010	Guardrails AI	6.6K ⭐	Python	18K	Validator Hub, reask self-correction loop	✅ Published
011	OpenAI Codex CLI	27K+ ⭐	Rust	549K	Queue-pair arch, Guardian AI approval, 3-OS sandbox	✅ NEW

Projects sorted by GitHub stars (descending).

🆚 Cross-Project Comparison

We maintain a detailed COMPARISON.md with side-by-side analysis across all 11 projects:

Quick Reference (creator, stars, language, type, license)
Architecture patterns (agent loop, extensibility, deployment)
Multi-agent & orchestration approaches
Memory & persistence strategies

We also have a new CROSS-CUTTING.md — a deep horizontal analysis covering:

Architecture pattern taxonomy across all projects
Context management strategies compared
Tool system design patterns
Security ratings (A-F grading)
Anti-pattern catalog with evidence
"Build from scratch" recommendations
Safety & security postures
Design patterns and anti-patterns found
"If I Were Building an Agent Today" synthesis

➡️ Read the full comparison →

🎯 What Makes This Different

Other "awesome" lists	This repo
Link to repos	Dissect repos
"What they use"	Why they chose it
Stars and badges	Architecture diagrams + trade-off analysis
Surface-level	Staff-engineer-level depth
Praise everything	Find real problems

📐 Each Teardown Includes

📄 README.md          — Full source code analysis + architecture decisions
⚠️ Problems found     — Bugs, code smells, architectural risks
🆚 Comparisons        — How it stacks up against alternatives

Coming Next

OpenCode (139K stars) — the open-source coding agent everyone uses
Cursor — the most controversial AI IDE
Windsurf — Codeium's AI IDE challenger

🤝 Contributing

Found an error? Have a better analysis? PRs welcome!

🐛 Fix factual errors
🔍 Add missing architecture details
💡 Suggest projects to teardown → Open an Issue
🌍 Help translate teardowns

📌 Stay Updated

New teardown every week. We're working through the top AI agent projects — Cursor, Windsurf, and others are on the list.

⭐ Star this repo → GitHub will notify you when new teardowns drop. That's literally what the star button does — it's a subscribe button, not a vanity metric.

👁️ Watch → Get notified for every new teardown, not just the ones that trend.

We've published 11 deep-dives so far. The next batch is already in progress.

💬 Discord: Join the community — discuss architecture decisions

⚠️ Disclaimer

These teardowns are independent technical commentary for educational purposes. All code snippets are cited under fair use for criticism, commentary, and education. Project names and trademarks belong to their respective owners. Star counts and LOC figures are snapshots at time of writing and may have changed.

If you're a project maintainer and have concerns about any content, please open an issue.

11 projects dissected. Updated weekly.
If you build AI agents, you should know how the best ones actually work.

awesome-ai-anatomy