What’s new
- guides
The latest guides, and what each is for.
Every guide here was published in the last few weeks. They circle one shift: you are not running a single agent any more, you are running several — and the bottleneck moved from writing code to reviewing it and knowing what your agents are actually doing.
Every guide is available in English and Polish.
Recently published
-
Level 5: You Run the Software Factory
The top rung of the free Ladder section: what the six stations of a software factory — intent, harness, loop, graph, verification, release — actually look like in Claude Code, Codex and Cursor.
Reach for it when agents are opening more pull requests than anyone can read, and you have to decide which parts of the pipeline are safe to stop watching.
Read the guide -
The State of Agentic Engineering, August 2026
Third-party evidence, dated: how much production code agents write in 2026, the autonomy ladder from assisted coding to the dark factory, and what stays human. Every number carries a publisher and a date.
Reach for it when someone asks what share of your code agents should write, and you need numbers with sources instead of a vibe.
Read the guide -
Compound Engineering: The Loop, the Plugin, the Evidence
Every's method for making each unit of work leave the system smarter — and an audit nobody else ran: both directories the guide points at are gone. Plus the measurement that argues its premise backwards.
Reach for it when you keep re-solving problems your team already solved, and nothing you write down ever gets read again.
Read the guide -
Anti-Slop Writing Skills: Unslop, miodkuj, and What the Evals Say
The anti-slop genre in one pass: what unslop bans, why its scanners run in CI without an agent, its own benchmark recording a no-ship — plus miodkuj, for the Polish tells no English word list reaches.
Reach for it when a reviewer keeps writing "this reads like ChatGPT" and nobody can point at the sentence that did it.
Read the guide -
Software Factories: How to Build One and Keep It From Rotting
Dex Horthy says factories fail, PostHog says they work, and both are right about different things. The ceiling is measured; the wound below it is fixable.
Reach for it when your agents open more pull requests than the team can review, and you have to decide what it is safe to stop reading.
Read the guide -
Dr. Skill: Audit What Your Agent Actually Loads
A read-only CLI that resolves what each coding agent has actually loaded — skills, MCP servers — then checks the result against 34 issue categories, from shadowed names to secrets sitting in configs.
Reach for it when your agent reaches for the wrong tool and you cannot tell what it actually has loaded — or before you let an unfamiliar MCP server near a repo.
Read the guide -
Pruning CLAUDE.md and AGENTS.md: The Ablation Protocol
The advice inverted in July: stop adding a rule after every mistake, start deleting. What each tool really loads and when, the truncation Codex performs without a word, and the prompts that do the cutting.
Reach for it when your context file has been growing for months and the agent still ignores the rule you added back in April.
Read the guide -
Alternative IDEs and Agent Shells
Everywhere you can run Claude Code, Codex and Cursor Agent — Zed and JetBrains over ACP, Conductor, Sculptor, Warp — with a decision table and one rule: does it run your agent, or its own?
Reach for it when you are choosing where to run agents and do not want to build a workflow on a tool that is gone in six months.
Read the guide -
herdr: The Agent Multiplexer With a Scriptable API
Real PTY panes and a persistent server like tmux, but every pane classified idle, working or blocked. Under the TUI sits a socket API: one agent can spawn, prompt and supervise the rest.
Reach for it when you are running six agents and one has been sitting on a permission prompt since you went for coffee.
Read the guide -
tmux for Agent Fleets: Sessions, send-keys, wait-for
Your agents create and clean up their own worktrees now, so the old tutorials sit a layer too low. What is left: sessions, send-keys between agents, wait-for barriers, and hooks that report exact state.
Reach for it when you need runs to survive a closed laptop or a dropped SSH connection, and you want agents to hand work to each other.
Read the guide -
Mobbin MCP: Real Design Examples for AI Agents
Ask an agent for a paywall and you get the statistical average of every paywall in its training data. This connects it instead to screens and flows captured from shipped iOS and web products.
Reach for it when your AI-generated interface looks like everyone else’s AI-generated interface.
Read the guide -
Taste Skill: The Anti-Slop Frontend Skill Library
How Leonxlnx/taste-skill constrains AI frontend generation up front: brief inference, three dials, a design-system map and a hard ban list — plus per-agent install and how it composes with the design pipeline.
Reach for it when every screen your agent generates arrives as the same centred card with a gradient button, and you would rather constrain the generation than correct the output.
Read the guide -
Superpowers: A Disciplined Design-to-PR Agent Workflow
How a plugin turns brainstorming, worktrees, executable plans, TDD, subagent review and fresh verification into one enforced workflow — with every installation target, the evidence behind it, and honest limits. Verified against upstream v6.1.1.
Reach for it when your agent runs are improvised from scratch each time and you want one workflow it cannot quietly skip a step in, from the first sketch to the pull request.
Read the guide -
What Claude Code Can Do That You Haven’t Tried
The underused Claude Code features from Gui Ferreira's NDC AI 2026 talk: one-time setup, your daily loop, extensibility, and work beyond code. Every command verified against the docs, not against the talk.
Reach for it when you have used Claude Code for months and suspect you are still driving it the way it worked a year ago.
Read the guide
Looking for everything else?
The documentation changelog records every new guide, lesson and content refresh, month by month, back to February 2026.
The Ladder section is free. The rest is one subscription.
Seven days free, then $99.99 a year: every guide in English and Polish, the four books, and the Setup Pack. Every guide shows the date it was last updated, so you can tell how old the advice is before you follow it.