Informal hub · not an official Anthropic page

Boris Cherny

Creator of Claude Code at Anthropic — frameworks, workflows, and team patterns collected here so you can start at mikesblogdesign.com/boris and go deep.

← AI Agents  ·  AI  ·  Productivity  ·  @bcherny on X

Core (Directly From Boris)

Framework

4 Steps of AI Adoption

Maturity ladder from Gated (0) to AI-native (4). Bottlenecks, guardrails, verification, and measuring real eng return — not just tokens.

Read → Philosophy

Loops, Not Prompts

Boris on Claude Code at Anthropic: stop one-shot prompting and write autonomous agent loops — the next abstraction layer.

Read → Teams

The 5 Roles in AI-First Companies

Prototyper, Builder, Sweeper, Grower, Maintainer — archetypes that cut across job titles on AI-first teams.

Read → Workflow

How Boris Uses Claude Code

His personal setup and practical habits for shipping with Claude Code (from his vanilla-but-disciplined thread).

Read → Config

CLAUDE.md Template

Project instruction patterns shared from Anthropic-style Claude Code team usage.

Read →

Loops Series & Practice

The loops series expands the “write loops, not prompts” idea into tutorials and checklists. Not every page is authored by Boris; all sit on the same adoption ladder.

Building AI Agent Loops & Workflows

Practical loops: cron + judgment, skills, agent-friendly CLIs — implementing the “loops not prompts” idea.

Read →

6-Month Path to Agentic AI Engineer

Roadmap from foundations to production multi-agent systems (loops series companion).

Read →

Stop Babysitting Your Agents

Verification loops, parallel agents, self-improving skills, background routines — Step 2→3 territory.

Read →

Claude Code (overview)

Getting started with Claude Code in the editor / CLI ecosystem.

Read →

Claude Skills

Package reusable expertise so agents become specialists (pairs with Step 3 standards).

Read →

Agentic Setup Checklist (0–18)

Harness inventory: router docs, worksheets, multi-model review, night shift — guardrails for each adoption step.

Read →

AI-Native Video Editing

Claude Code skills orchestrating FFmpeg + Hyperframes for same-day video post — Step 2–3 style human gates.

Read →

Related Context

AI Questions Executives Are Asking

Field questions about security, spend, skills, and “are we behind?” — often Step 0–2 org pain.

Read →

Distribution Engineer

Growth as infrastructure; connects to AI-first team roles.

Read →

Harness: AI Coding Tips

Day-to-day agent coding practices that support Step 1–2.

Read →

Chiang Mai Claude Code Meetup

Local community notes with links into the loops series.

Read →

Suggested Reading Order

  1. 4 Steps of AI Adoption — where is your team on the ladder?
  2. Loops, Not Prompts — why the abstraction shifts at Step 2+
  3. 5 Roles — who does what when the org scales AI
  4. How Boris uses Claude Code — personal Step 2+ habits
  5. Stop Babysitting Agents + Agentic Setup — guardrails and harness
  6. CLAUDE.md + Skills — encode standards for Step 3

Unofficial collection of notes and frameworks related to Boris Cherny’s public writing on Claude Code and AI teams. Always verify product names and features against current Anthropic docs.

Comments

Approved comments appear below. Log in once with GFAVIP — it applies across the whole site. GFAVIP login

View comments archive