Anthropic's Skill authoring best practices
platform.claude.com/docs/en/agents…

System Design & Architecture List
generativeprogrammer.com/lists/system-d…

System design is more than drawing boxes and arrows. It's about understanding trade-offs and making better engineering decisions.

Here is good reading list for any software engineer.

The Book of Secret Knowledge feels like a senior operator’s bookmarks turned into a field manual.

Inside: CLI tools, systems and network references, containers, pentesting resources, shell tricks, and cheat sheets.

240K stars.

github.com/trimstray/the-…

🌟 Claude Code Mods Explained 🌟
↳ a programmable extension layer for your coding session (like stateful hooks).

Instead of only prompting the agent, you can run JavaScript modules directly inside the session to customize its behavior.

How it works:
↳ Mods run inside your eact automatically to agent events
↳ Observe, rewrite, or intercept tool execution
↳ Maintain state across turns
↳ Render custom, interactive UI

Why it matters:
↳ Stateful workflows across multiple turns
↳ Deterministic behavior without relying on prompts
↳ Custom guardrails, automation, and developer tooling

Claude Code is becoming more than a coding assistant. It's becoming a programmable agent runtime.

claude.dev/blog/getting-s…

The 80/20 rule for your whole life
yewjin.substack.com/p/always-be-cu…

How Uber is using AI agents to delete code at scale.

↳ 16,000+ stale feature flags removed
↳ 1M+ lines of code cleaned up
↳ Automated rewrites, CI fixes, and validation

uber.com/us/en/blog/ai-…

Nik Kale on operational MCP security:

→ CI gates for tool execution
→ Isolate tooling and scope tokens
→ Diff-based manifest reviews

Read on @InfoQ: Securing MCP in Production: Defense-in-Depth beyond the Gateway
infoq.com/articles/secur…

A Proposed Framework For Evaluating Skills - @tessl_io
tessl.io/blog/a-propose…

These are the AI newsletters I regularly read and recommend.

I've collected them all in one place.

🔖 Bookmark the list: generativeprogrammer.com/lists/ai-a-team

Lalit Maganti's problem-finding loop:

Solve one real problem → earn trust → enter earlier conversations → see more of the org → spot the next pattern sooner.

A common shape is only a hypothesis. Pressure-test it before building.
lalitm.com/post/find-prob…

👌 Hunger beats tenure: Advice for junior engineers
felixrieseberg.com/hunger-beats-t…

Developing Taste
emilkowal.ski/ui/developing-…

Zalando auto-approves PRs its internal rules classify as low risk. That covered 33% of its PRs and cut their lead time by 20% to 40% versus all PRs. The rules come from its own incident history.
engineering.zalando.com/posts/2026/08/…

Chain of Thought For Humans
boristhebrave.com/2026/10/18/cha…

The Brain Sandwich
terriblesoftware.org/2026/10/02/the…

The big idea: Think first. Use AI. Think again.

Made it to #69 on Substack's Rising Tech leaderboard!

Every fortnight, I share what I learn about AI, GenAI patterns, tools, and resources worth exploring.

📩 Sub and support generativeprogrammer.com

The ability to build anything has come with a critical lesson:

Engineers have really, really bad ideas.

How to Measure the Human Cost of Agent Work
cutler.sg/blog/2026-10-m…