Every guide, filed by section and by level.
577 guides in 12 sections, in English and Polish. Light up the guides that start at your level, filter by a word and open the one you need.
$59.99 once for 12 months of access. Nothing renews on its own. 30-day refund on a first purchase.
- Sections in the library
- 12
- Guides, in English and Polish
- 577
- Free to read: Start here and Autonomy Ladder
- 19
- Tools with a guide tree of their own
- 3
The whole library, drawn to scale.
One tick is one guide. A taller tick starts higher on the autonomy ladder. Pick a level to see where its guides sit.
Pick a level to light up its guides. The short grey ticks are guides that carry no level.
Level 1, assisted: mostly the quick starts for one tool.
Level 2, paired: the largest group in the library.
Level 3, review manager: testing and verification, review automation and shared team practice.
Level 4, spec manager: specs, test oracles and running agent-written code in production.
Level 5, dark factory: automation, SDKs, scheduled runs and agent fleets.
- Start here 8
- Autonomy Ladder 11
- Cursor 70
- Claude Code 79
- Codex 56
- Workflows & method 110
- Ecosystem & comparisons 116
- Cookbook 41
- Lead a team 24
- Engineering organization 40
- Strategy & business 11
- Reference & help 11
Open a section.
Sections are ordered by how often people use them. The Start here and Autonomy Ladder sections are free to read. In every other section you read 5 articles a month free, 10 with a free account, and all of them with the Toolkit.
Nothing matches that. Try a shorter word.
Start here
A reading path for each role, and the quick wins for your first day.
- Developer track: a reading path to agentic engineering, Free
- Tech lead track: move the whole team up one rung, Free
- CTO and VP Engineering track: an operating model for agent-built software, Free
Show all 8 guides in Start here
- Start here: pick your path through agentic engineering, Free
- Developer track: a reading path to agentic engineering, Free
- Tech lead track: move the whole team up one rung, Free
- CTO and VP Engineering track: an operating model for agent-built software, Free
- Executive track: the ten-minute brief and the five decisions you own, Free
- Why AI Coding Tools? From Autocomplete to Agentic Engineering, Free
- Quick Wins in Your First 24 Hours, Free
- Power-user tips for Cursor, Claude Code and Codex, Free
Autonomy Ladder
Where your workflow sits today, and the guide for the next rung.
- The Autonomy Ladder: Which Level Is Your Workflow At?, Free
- One map: the autonomy ladder, the lifecycle and the factory stations, Free
- Level 1–2: Assisted and Paired Coding, Free
Show all 11 guides in Autonomy Ladder
- The Autonomy Ladder: Which Level Is Your Workflow At?, Free
- One map: the autonomy ladder, the lifecycle and the factory stations, Free
- The State of Agentic Engineering, September 2026, Free
- Vibe Coding vs Agentic Engineering: 21 Terms Compared, Free
- Level 1–2: Assisted and Paired Coding, Free
- Level 3: You Review the Diffs, Free
- Level 4: You Write the Specs, Free
- Level 5: You Run the Software Factory, Free
- Reading evidence instead of code, Free
- The Human's Job in Agentic Engineering, Free
- Your craft and career when agents write the code, Free
Cursor
Rules, agent modes, the Agents Window, Bugbot and cloud agents, in nine task folders.
- Cursor quick start: from install to a verified first feature
- Project rules in Cursor
- The Agents Window in Cursor: running parallel agents
Show all 70 guides in Cursor
Overview 1
Quick Start 11
- Cursor quick start: from install to a verified first feature
- Install Cursor and migrate from VS Code
- Essential Cursor configuration for agent work
- Model selection in Cursor
- Project rules in Cursor
- Turn a PRD into a plan and tasks in Cursor
- Set up MCP servers in Cursor
- Context management in Cursor
- Build your first AI-assisted feature
- Debugging and Error Recovery
- Version control and team collaboration in Cursor
Plan 4
Build 23
- Build with Cursor: from agent chat to verified change
- Cursor core features: pick the right tool and verify the result
- Refactoring legacy code with Cursor
- Cursor Tab: get multi-line edits right and verify them
- Building API integrations with Cursor's agent
- Cursor inline edit: tips 46–60
- Agent chat workflows in Cursor
- Build frontend UI from designs with Cursor
- Generate and verify documentation with Cursor's agent
- Microservices development in Cursor
- Pair programming with Cursor's agent
- Mobile app development with Cursor
- Build a data pipeline with Cursor
- Framework migration in Cursor
- Cursor agent modes: when to plan, ask, build and debug
- Checkpoints and branching: undo Cursor agent changes safely
- Large codebase strategies in Cursor
- Multi-repo workflows in Cursor
- Prompt templates for Cursor's agent
- Refactoring strategies in Cursor
- Snippet management in Cursor
- Cursor time savers for everyday chores
- The Agents Window in Cursor: running parallel agents
Test & Debug 4
Review & Ship 6
Operate 6
Automate 7
- Cursor automation: pick the surface, then prove the result
- Build a custom MCP server with Cursor's agent
- Advanced Cursor techniques: tips 91-105
- Scripting the Cursor CLI: git hooks, CI review and scheduled jobs
- Cloud Agents and Automations: Cursor's Factory Floor
- Custom rules and templates in Cursor
- The Cursor SDK and the Cloud Agents API
Team & Enterprise 3
Claude Code
Memory, hooks, subagents, goals and routines, in the same nine task folders.
- Claude Code memory: CLAUDE.md, rules and auto memory
- Claude Code hooks: guardrails, gates, and live context
- Custom subagents in Claude Code: scoped specialists you can verify
Show all 79 guides in Claude Code
Overview 1
Quick Start 12
- Claude Code quick start: from install to a verified first feature
- Install Claude Code on macOS, Linux, and Windows
- Claude Code authentication: subscriptions, API keys, and cloud providers
- Configure Claude Code: settings, permissions, and models
- Claude Code IDE integration: VS Code, JetBrains, and any editor
- Initialize a project for Claude Code: CLAUDE.md, rules, and memory
- Build your first feature with Claude Code
- Version control with Claude Code: commits, conflicts, PRs and worktrees
- PRD to plan to tasks in Claude Code: a reviewable path from requirements to code
- Set up MCP servers in Claude Code
- Deep reasoning in Claude Code: effort levels and extended thinking
- Error recovery in Claude Code: rewind, reset, and get unstuck
Plan 5
Build 22
- Build with Claude Code: from an approved plan to a proved change
- What Claude Code Can Do That You Haven't Tried
- Claude Code command line: sessions, flags and one-shot pipes
- Large codebases and monorepos in Claude Code: scope, plan, verify
- IDE and CLI coordination: run Claude Code beside your editor
- Claude Code workflow optimization: queue, remember, codify, verify
- Efficiency hacks for daily Claude Code work
- Build REST and GraphQL APIs with Claude Code
- Batch operations: one change across hundreds of files
- Database work in Claude Code: schemas, migrations, queries and seed data
- Generate and maintain documentation with Claude Code
- From plan to working code with Claude Code
- Build third-party API integrations with Claude Code
- Claude Code memory: CLAUDE.md, rules and auto memory
- Database and code migrations with Claude Code
- Multi-file workflows
- Prompt engineering for Claude Code
- Large-scale refactoring with Claude Code: change structure, prove behaviour
- Refactoring patterns in Claude Code
- Claude Code slash commands and shortcuts, by job
- Terminal setup for Claude Code: tmux, notifications and shell pipes
- Claude Code Desktop: sessions, worktrees and review in the app
Test & Debug 3
Review & Ship 6
Operate 7
Automate 12
- Claude Code automation: pick the surface, then prove the result
- Advanced Claude Code techniques
- Agent view: run many Claude Code sessions with claude agents
- Scripting Claude Code: Headless Runs, Pipes, Cron and Git Hooks
- Custom commands and skills in Claude Code
- Custom subagents in Claude Code: scoped specialists you can verify
- Dynamic workflows and ultracode
- Goal workflows with /goal in Claude Code
- Claude Code hooks: guardrails, gates, and live context
- Routines: Scheduled, API and GitHub-Triggered Claude Code Runs
- The Claude Agent SDK: building your own agents on Claude Code's loop
- Channels and Remote Control: driving Claude Code sessions from anywhere
Team & Enterprise 7
Codex
AGENTS.md, config.toml, worktrees and codex exec, in the same nine task folders.
- Set up AGENTS.md for Codex
- Configure Codex with config.toml: model, permissions, and profiles
- Isolated Codex tasks with Git worktrees
Show all 56 guides in Codex
Overview 1
Quick Start 10
- Codex quick start: from install to a verified first task
- Install Codex: CLI, desktop app, and IDE extension
- Authentication
- Configure Codex with config.toml: model, permissions, and profiles
- Run your first Codex task
- Set up AGENTS.md for Codex
- Connect Codex to GitHub
- Review Codex's changes before you commit
- Set up MCP servers in Codex
- Recover from Codex errors and wrong turns
Plan 4
Build 13
- Build with Codex: context, isolation and gates
- Optimize AGENTS.md and skills for Codex
- Build APIs with Codex: contract first, verified by machine
- Codex desktop app: parallel threads, worktrees and review
- Codex batch operations: one change across many packages
- Codex CLI commands, slash commands and shortcuts
- Legacy code modernization at scale with Codex
- Context Management Across Codex Surfaces
- Database work with Codex: schemas, migrations, and queries you can prove
- Generate and maintain documentation with Codex
- Codex efficiency hacks: configure once, save time every session
- Prompt Codex with task briefs, plan mode and AGENTS.md
- Large-scale refactoring with Codex worktrees
Test & Debug 3
Review & Ship 6
Operate 5
Automate 9
- Codex automation: pick the surface, then prove the result
- Advanced Codex setups: providers, sandboxes and audit trails
- Scheduled Codex automations you can trust unattended
- Codex cloud environments: setup, caching, network and best-of-N
- Build with the Codex SDK: threads, structured output and verification
- Codex multi-agent workflows: parallel lanes, subagents and cloud fan-out
- Run Codex non-interactively with codex exec
- Codex in Slack and Linear: delegate issues to cloud tasks
- Isolated Codex tasks with Git worktrees
Team & Enterprise 3
Reference 2
Workflows & method
The method that holds in all three tools: artifacts, context, the harness, verification and shipping.
- The artifact chain
- Spec-driven development: the spec as the source of truth
- How strong is your oracle? Trusting tests you did not read
Show all 110 guides in Workflows & method
Overview 3
The lifecycle 10
- The AI-native software development lifecycle
- The artifact chain
- Plan: capture intent.md
- Design: write spec.md
- Build: plan.md, then implement
- Test: give the session a feedback loop
- Deploy: review in both directions, gate production
- Maintain: close the loop
- Metrics for an AI-native SDLC
- Tool map: Claude Code, Cursor, and Codex
Principles 8
- Principles for working with coding agents
- Software Factories: How to Build One and Keep It From Rotting
- Human in the loop: judgment at the gates, not in every diff
- Test-driven development with AI agents
- Error-Driven Development
- Continuous Delivery with AI Assistance
- When to use agent mode vs ask mode
- Effective prompting techniques for coding agents
Context 9
- Context management for coding agents
- Context windows and token limits for coding agents
- Using Documentation as Effective AI Context
- Project structure optimized for AI coding agents
- Code search and indexing for AI agents: text, semantic and symbol search
- Long-Term Context Retention Patterns
- Prune CLAUDE.md and AGENTS.md: the ablation protocol
- Context cost optimization for coding agents
- AGENTS.md and CLAUDE.md — concise repository context
Harness 8
- The harness: everything the agent runs inside
- Permissions, sandboxes and approval modes across Claude Code, Codex and Cursor
- Making a codebase agent-ready
- Ephemeral environments: a full stack per agent task
- Use hooks as deterministic guardrails
- Agent skills — turn policy into a tested workflow
- Scoped subagents — delegate bounded work, not responsibility
- Run parallel agents in isolated worktrees
Plan & specify 10
- Plan and specify: the contract before the code
- Spec-driven development: the spec as the source of truth
- Executable acceptance criteria: from story to failing test
- Architecture decisions agents can follow
- Four-artifact design-to-code pipeline
- Evidence-based rating and iteration
- Behavior-driven development with AI agents
- Domain-driven design with AI coding agents
- Agile Workflows: Scrum and Kanban Integration
- Visual acceptance before frontend implementation
Build 7
- Build with agents: the workflow for your codebase
- Million+ LOC strategies for coding agents
- Monorepo Workflows with AI Assistants
- Legacy modernization with agents: a program, not a prompt
- Distributed Systems Development with AI
- API development with AI: contract first, verified by machine
- Database Design and Queries
Test & verify 22
- Testing and verification for agent-written code
- How strong is your oracle? Trusting tests you did not read
- Protecting the oracle: checks the agent cannot edit
- Unit test strategies and generation with AI agents
- Integration Test Patterns
- API Test Automation
- End-to-End Test Automation
- Property-based testing for agent-written code
- Model-graded checks: rubric judges where tests cannot reach
- Architecture fitness functions: maintainability without reading the code
- Catching plausible-but-wrong code before merge
- Hallucinated packages and slopsquatting: dependency checks for agent changes
- Characterization tests: pin legacy behavior before agents touch it
- Code quality gates for agent-written code
- Accessibility testing with AI agents: WCAG 2.2 AA gates in CI
- Load, Stress, and Benchmark Testing
- Security testing for agent-written code
- Fault injection testing: prove timeouts, retries, and circuit breakers fire
- Mobile app testing with AI agents
- Test data strategies and fixtures
- Continuous evals — regression-test the agent harness in CI
- Agent-driven browser verification — prove the user flow
Review & ship 11
- Review and ship
- The evidence bundle: what an agent's pull request must prove
- Reviewing an agent's pull request without reading every line
- Pipeline automation with AI
- Progressive delivery for agent-written changes
- Docker and Kubernetes Containerization
- Infrastructure as code with AI agents
- Configure review agents for layered pull-request review
- Run a bounded PR review-fix loop
- Production approval as a system boundary
- Rehearsed rollback with human production authority
Operate & learn 14
- DevOps with AI
- Observing the agents: traces, tool calls, cost and audit logs
- Keeping an agent-written codebase healthy
- A failure taxonomy for agent-written changes
- Monitoring and observability
- AI-powered incident response
- Production Performance Optimization
- Security Operations Automation
- Regulatory Compliance Automation
- AI-Powered Disaster Recovery
- Cloud cost management and FinOps with coding agents
- Deterministic anomaly detection and scoped diagnosis
- Scheduled security scans with evidence and PR gates
- Scoped AI incident triage
Automate 8
- Automate
- The /goal Command: Goal-Directed Autonomous Runs
- The /loop Command: Recurring and Self-Paced Prompts
- Background and cloud agents compared
- From issue to pull request with no hands on the keyboard
- From intent to production: a worked end-to-end pipeline
- Driving agents from code: Claude Agent SDK, Codex SDK and Cursor SDK
- Multi-agent orchestration patterns: planner, workers, judge
Ecosystem & comparisons
MCP servers, skills, plugins, agent tools and frameworks, plus the comparisons between tools.
- Cursor vs Claude Code: when to use which
- Cursor vs Claude Code vs Codex: feature comparison matrix
- The MCP starter stack for 2026: five servers most teams need
Show all 116 guides in Ecosystem & comparisons
Choose a tool 20
- Tool Comparison Overview
- The coding-agent landscape beyond the big three
- GitHub Copilot as a coding agent: cloud agent, app, CLI and billing
- Gemini CLI, Jules and Antigravity
- OpenCode: the open-source agent for any model
- Kiro: spec files, property-based tests and Kiro Crew
- Cursor vs Claude Code: when to use which
- Codex vs Cursor and Claude Code: strengths and trade-offs
- Claude Code, Codex and Cursor vs GitHub Copilot
- Cursor, Claude Code, and Codex vs Devin Desktop (formerly Windsurf)
- From plain chat to integrated development
- Cursor vs Claude Code vs Codex: feature comparison matrix
- AI coding tool plan prices and how usage is billed
- Moving to Cursor, Claude Code, or Codex: a migration guide
- Adding Claude Code or Codex alongside GitHub Copilot
- Migrating from Devin Desktop (formerly Windsurf)
- Migrating between Codex, Cursor, and Claude Code
- Migrating from traditional IDEs
- Choose a primary AI engineering harness
- Size an AI plan from measured workload
Models 5
MCP servers 31
- Introduction to Model Context Protocol
- The MCP starter stack for 2026: five servers most teams need
- Top community MCP servers and how to vet them
- Top 30 MCP Servers for Software Teams, Ranked (September 2026)
- MCP best practices: context budget, speed and parallel agents
- Securing MCP Servers: Scanning, Read-Only Modes and Prompt Injection
- MCP registries and gateways: allowlisting servers for an organization
- MCP 2026-07-28: what changed and how to migrate a server
- Building your own MCP server
- Reducing MCP token cost: Code Mode, context-mode and CLI-plus-skill routes
- MCP Server Connection Issues
- Atlassian Rovo MCP: Jira, Confluence and Bitbucket from Your Agent
- Playwright MCP vs Playwright CLI: Browser Testing for Coding Agents
- Browser Automation for Agents: Playwright, Chrome DevTools, Browser Use, Stagehand and agent-browser
- Web Search and Scraping MCP: Firecrawl, Exa, Tavily, Brave and MarkItDown
- Cloudflare MCP: Code Mode and the Domain Servers
- Debug Production from the Editor: Sentry, Grafana, Datadog, and PostHog MCP
- Database MCP Servers: Postgres MCP Pro, DBHub, MCP Toolbox, Supabase, Neon, and MongoDB
- Figma MCP and Design-to-Code in 2026
- AI Developer Toolkit MCP Server
- Context7: Current Library Docs for Your Agent (MCP, CLI and Skills Mode)
- The Reference MCP Servers and When You Don't Need Them
- Cloud and Platform MCP Servers: AWS, Azure, Google Cloud, Vercel and Stripe
- Infrastructure MCP: Docker MCP Toolkit, Kubernetes and Terraform
- Mobbin MCP: Real Design Examples for AI Agents
- Tickets and Docs over MCP: Linear, Jira, Notion and Slack
- Dev-Server MCPs: Next.js DevTools, MobileBuildMCP and Hugging Face
- One server, many apps: Composio and n8n-mcp
- shadcn/ui MCP Server
- GitHub MCP Server: Remote, Local, Toolsets and Lockdown Mode
- MCP Servers That Understand Your Codebase: Serena, codebase-memory-mcp, Claude Context and Repomix
Agent skills 20
- Introduction to Agent Skills
- Installing, sharing and pinning skills across a team
- Building and Publishing Your Own Skills
- Skill Supply-Chain Security
- Dr. Skill: Auditing Your Agent's Skill and MCP Loadout
- The 25 Most-Installed Development-Practice Skills (September 2026)
- Skills That Make Agents Test and Debug Properly
- Code Review Inside the Agent: Skills and Plugins
- Documents and Writing: Office Skills, doc-coauthoring and writing-guidelines
- Skills for Context and Token Discipline: caveman, handoff, karpathy-guidelines
- agent-browser: letting the agent check the running app
- Official Vendor Skills: Who Publishes What (September 2026)
- Best Frontend Skills: Design, React, Next.js and Mobile
- Best Backend and DevOps Skills: Databases, Auth, Cloud, IaC
- Skills for Product, Marketing and Knowledge Work
- Matt Pocock's Skills: From grill-me to implement
- Impeccable: a design vocabulary for your AI harness
- shadcn/improve: Audit With Your Best Model, Execute With a Cheap One
- Taste Skill: The Anti-Slop Frontend Skill Library
- Anti-Slop Writing Skills: Unslop, miodkuj, and What the Evals Say
Plugins 7
- Plugins and marketplaces in Claude Code, Codex and Cursor
- The plugins worth installing, with usage examples
- Anthropic's Developer Plugins: code-review, feature-dev, pr-review-toolkit, security-guidance and More
- Marketplaces and registries: where plugins, skills, and MCP servers come from
- Persistent Memory for Agents: claude-mem, planning-with-files and MCP Memory Servers
- Building and distributing a plugin or a private marketplace
- Publishing a Private Plugin Marketplace for Your Team
Agent tools 11
- Agent tools: multiplexers, desktop environments, sandboxes, gateways and more
- Desktop agent environments compared: Orca, Conductor, Superset, Emdash, Warp and more
- Running 10 Agents at Once with herdr, Conductor and Worktrees
- herdr: The Agent Multiplexer with a Scriptable API
- tmux for agent fleets: sessions, send-keys, wait-for
- Agent sandboxes compared: container-use, E2B, Daytona, microVMs and Docker
- Headless Agents in CI: claude -p, codex exec and the Agent SDKs
- Beyond the Big Three: OpenCode, Pi, Copilot CLI, Cline, Goose and Other Coding Agents
- Driving Agents from Your Phone: Remote Control, Happy, CloudCLI and Orca Mobile
- Gateways and Local Models: LiteLLM, CC Switch, Claude Code Router, Ollama and LM Studio
- Voice Input for Agentic Coding
Frameworks 15
- Agentic development frameworks compared
- Superpowers: A Disciplined Development Workflow for Coding Agents
- Spec-Driven Frameworks Compared: Spec Kit vs OpenSpec vs BMAD vs Superpowers
- GitHub Spec Kit: spec-driven development in practice
- Compound Engineering: The Loop, the Plugin, the Evidence
- The BMAD Method: agile roles as agents
- OpenSpec: change proposals and living specs
- GSD: context-engineered phases for solo builders
- gstack: a role-based skill stack for Claude Code
- Everything Claude Code: the largest bundle, used for real
- Discipline packs: Pocock skills, gstack, agent-skills, Ponytail and ECC
- Hours of Autonomy: GSD Core, Ralph Loops, /goal and oh-my-claudecode
- Multi-Agent Harnesses and Catalogues: Ruflo, wshobson/agents, SuperClaude, Task Master and CCPM
- Frameworks with audit trails and standards: AI-DLC, Agent OS, Tessl and Kiro-style specs
- One Rules Source for Every Agent: AGENTS.md, Ruler and rulesync
Observability & evals 6
- Measuring Agentic Engineering: Telemetry, Cost, Review, Security and Evals
- What Did the Agents Cost? ccusage, /usage, Monitors and Gateway Budgets
- Security Gates for Agent-Written Code: Semgrep, Gitleaks, TruffleHog, Snyk and Claude Security Review
- OpenTelemetry and Analytics for Claude Code, Codex and Cursor
- AI Code Review Bots Compared: Claude Code Review, Codex, Bugbot, CodeRabbit, Greptile and More
- Evaluating Agents, Prompts and CLAUDE.md Changes: promptfoo, Inspect SWE, Langfuse and Braintrust
Cookbook
Prompts for your stack, each paired with the check that proves the result.
Show all 41 guides in Cookbook
Overview 1
Frontend 6
Backend 12
Database 5
Data 1
DevOps 6
Mobile 6
Error handling 4
Lead a team
Rolling agents out to a team: the shared harness, the review flow and the people.
- Rolling out coding agents to a team
- Helping the team stop reading every diff
- Running the review queue when agents open the pull requests
Show all 24 guides in Lead a team
Roll out 7
- Rolling out coding agents to a team
- Adoption roadmap for one repository: ladder transitions per loop
- Team onboarding and adoption, loop by loop
- Project conversion playbook
- Workflow transformation: rebuilding feature delivery around agents
- Team adoption — measure effective workflow use
- Developer onboarding — prove one safe workflow
Shared harness 5
Review & flow 5
Engineering organization
The operating model, security and compliance, measurement, and spend.
- The agent threat model: prompt injection, data and blast radius
- An AI usage policy engineers will follow
- The EU AI Act for companies building software with agents
Show all 40 guides in Engineering organization
Overview 4
Autonomy & controls 8
- Governance and autonomy: put humans at the gates
- Plan policy: accepted artifacts before risky execution
- E2E policy: browser evidence for every user-visible change
- AI in CI/CD: a bounded server-side loop
- Unattended agent runs at Level 4–5 — bounded recurring work
- Change provenance and risk routing
- Prototype policy — graduate by risk and evidence
- Enforcing one policy across every coding agent you run
Security, data & compliance 13
- The agent threat model: prompt injection, data and blast radius
- Agent identity, credentials and secrets
- Security standards and compliance for coding agents
- MCP security — authorize each identity and tool
- Data privacy and enterprise policies for coding agents
- Running coding agents on a corporate network
- An AI usage policy engineers will follow
- The EU AI Act for companies building software with agents
- CRA, NIS2 and DORA-EU: EU software regulation for teams using agents
- Agentic engineering in regulated industries
- Legal and IP questions about agent-written code
- ISO/IEC 42001, NIST AI RMF and SOC 2 evidence from the agent pipeline
- When an agent causes an incident
Measure outcomes 5
Spend & vendors 8
- AI usage cost governance
- Team accounts — ownership, lifecycle, and data terms
- Tooling policy — approved workflows and measured exceptions
- AI tooling roadmap — the portfolio-item template for capability bets
- Vendor risk management — test continuity, not logos
- Where the model runs: gateways, cloud routes, residency and zero retention
- Avoiding lock-in: portable context, skills and pipelines
- Buying AI coding tools: the security, legal and data questionnaire
Strategy & business
The business case, the economics, board reporting and how teams change shape.
- Building the business case for agentic engineering
- The economics of agent-built software: cost per accepted change
- Reporting AI engineering to the board
Show all 11 guides in Strategy & business
- Strategy: AI-native engineering for business leaders
- The economics of agent-built software: cost per accepted change
- Building the business case for agentic engineering
- AI coding success stories: the case studies that hold up
- How teams, roles and headcount change shape
- The organization-wide roadmap: from pilots to an agentic delivery system
- Leading people through the change
- Reporting AI engineering to the board
- AI-native delivery for software houses and agencies
- Product management when build time collapses
- Software built outside engineering
Reference & help
Command and config reference, the glossary, troubleshooting and what changed.
- Command Reference Cheatsheet
- Configuration reference
- Keyboard shortcuts for Claude Code, Codex and Cursor
Show all 11 guides in Reference & help
Reference 5
The same nine tasks in every tool.
Cursor, Claude Code and Codex are each filed under the same nine task folders, from quick start to reference. The numbers are guides per folder.
| Task folder | Cursor | Claude Code | Codex | What is in it |
|---|---|---|---|---|
| Quick Start | 11 | 12 | 10 | Install, authenticate, configure, run a first task, set up MCP |
| Plan | 4 | 5 | 4 | Architecture, new projects, codebase analysis |
| Build | 23 | 22 | 13 | Features, refactoring, APIs, databases, documentation |
| Test & Debug | 4 | 3 | 3 | Debugging, test generation, test-driven development |
| Review & Ship | 6 | 6 | 6 | Code review, CI/CD, deployment |
| Operate | 6 | 7 | 5 | Cost, monitoring, performance, security audits |
| Automate | 7 | 12 | 9 | Headless runs, SDKs, subagents, scheduled automations |
| Team & Enterprise | 3 | 7 | 3 | Shared configuration, governance, privacy and security |
| Reference | 5 | 4 | 2 | Changelog, upgrades, shortcuts |
| Tool overview | 1 | 1 | 1 | The hub page for the tool |
| Guides in the tree | 70 | 79 | 56 |
Not sure where to start?
Four ways in, each one a page you can open now.
Questions about the library.
Something else? Pick a reason on the Contact page and write to the right address.
Can’t I just ask my agent instead of reading guides?
You can, and the answer will often sound right. Models invent flags: ask for Codex approval settings and you may get --approval-mode, which is not a Codex flag. The real one is --ask-for-approval (Codex CLI 0.159.2, checked 2026-10-02). Every guide shows when it was last updated, so you can see what still holds.
Can I read a guide before I buy?
Yes. The Start here and Autonomy Ladder sections are always free: the reading path for your role, the quick wins for your first day, and the Ladder section's 11 guides: each level named, how to test for it and how to climb to the next. Every other section gives you 5 full articles a month free, 10 with a free account, and all of them with the Toolkit.
How do I know a guide is current?
Every guide shows when it was last updated, and many name the tool version they were checked against, such as the hooks guide (Claude Code v2.1.283).
I only use one of the three tools. Is most of this wasted?
Each tool has its own tree, filed under the same nine task folders from quick start to reference. The sections on workflows, the cookbook and organizations describe the same method for all three.
Is the Polish library the same?
Yes. The English and Polish libraries list the same pages. Product and file names such as CLAUDE.md and AGENTS.md stay in English.
What do the levels on this page mean?
They come from Dan Shapiro’s five levels, from spicy autocomplete to the dark factory (“The Five Levels” (opens in a new tab), 2026-01-23). A guide’s level is the rung at which it starts being useful.
Engineers at these companies work with AI Developer Toolkit
- Grupa Morizon-Gratka
- Ringier Axel Springer
- TasteRay
Creator of wondel.ai Skills — 2.3k+ stars on GitHub (opens in a new tab)
Does this page answer your questions?
One tap, no sign-in.
What are you here for?
What is holding you back?
How does it read?
Thanks. If you left an email, Michał will reply.
Answering your questions?
One tap each. Close whenever you like.
no sign-in
Read the guide you came for.
Open any of the 577 guides with the Toolkit, in English or Polish.
$59.99 once for 12 months of access. Nothing renews on its own. 30-day refund on a first purchase.
Michał replies, not a bot. The address is used for this reply only.