📖ai-agent-book
AI Agent PDF
⭐ Very popular: 38k stars, gaining about 4k a week
View on GitHub ↗repo profile
momentum
durability
bus factor = how many people it takes to cover more than half the commits (6 months). 1 is a solo project; higher means the work is spread across a team. top-author share is the single busiest author's slice of those commits.
since we covered it
why it's a big deal
- It open-sources the harness craft. The book’s thesis is that everything around the model - context engineering, tool design, evaluation - is where the real competitive edge lives, and nearly every claim comes with an ablation experiment you can run yourself instead of a slide you have to believe.
- It treats agent engineering as a discipline with a curriculum: context engineering, memory and RAG, tools and MCP, coding agents, evals, post-training, self-evolution, multimodal, multi-agent. Each chapter is a lab, not a lecture.
- It is documentation-as-code at book scale: full text and figures under an open license, figures generated by scripts, errata and new chapters arriving as pull requests. The knowledge frontier is publishing in public, and increasingly not in English first.
under the hood
- Ten chapters plus introduction and afterword in Markdown, compiled to PDF via pandoc and xelatex; every diagram is generated by a Python script checked into the repo.
- Companion projects per chapter: prompt-injection attack/defense matrices, KV-cache-friendly context design, agentic RAG pipelines, MCP tool servers, a production-grade coding agent with 17 tools, and eval harnesses wired to GAIA, SWE-bench and Terminal-Bench.
- Projects come in three honesty tiers - runnable out of the box, reproduction guide, or design doc - with external benchmarks and training frameworks pulled in via a clone script. Apache-2.0.
our take from PR#39, 2026-07-30
star history
- PR#39 27k 2026-07-30
- now 38k + 11k since first covered
curve is sampled from GitHub's star history, plus our own daily readings since we covered it; the dashed stretch is before we first covered it, the solid line since. figures at coverage are the numbers we printed then (approx.), current count is live.
understory
Better known than its recent output, coasting a little on attention.
- output, commits & releases
- clout, star velocity
output = commits & releases; clout = star velocity, both 0 to 100 monthly indices; the gap where output runs above clout is the understory. The understory →
covered in
-
an agent-engineering degree in one repo
similar projects
compare these →- 📚 Haystack
Open-source AI orchestration framework for building context-engineered, production-ready LLM applications. Design modular pipelines and agent workflows with explicit control over retrieval, routing, memory, and generation. Built for scalable agents, RAG, multimodal applications, semantic search, and conversational systems.
26k ACTIVE - 🗜️ headroom
Compress tool outputs, logs, files, and RAG chunks before they reach the LLM. 20% fewer tokens for coding agents, 60-95% fewer tokens for JSON, same answers. Library, proxy, MCP server.
66k ACTIVE - 📈 Vibe-Trading
"Vibe-Trading: Your Personal Trading Agent"
31k ACTIVE
comments
Sign in with GitHub to add your blip on ai-agent-book.