Kieran Klaassen

Overview

Kieran Klaassen is presented in the newsletter as a creator demonstrating advanced workflow patterns for Claude Code, centered on his Compound Engineering plugin. His system shows how AI-assisted software development can become more reliable over time by combining planning, coding, assessment, and codification into a repeatable loop. Rather than treating each AI coding session as isolated, Klaassen’s approach captures what worked into local documentation so future runs inherit those learnings.

For AI Product Managers, this matters because it turns AI coding from a one-off productivity trick into an operational system. Klaassen’s workflow illustrates how teams can structure prompts, testing, documentation, and feedback loops so generated output improves across iterations. The examples highlighted in the newsletter also show how these patterns support unattended execution, persistent project memory, and automated browser testing—capabilities that are increasingly relevant for PMs managing AI-enabled product development.

Key Developments

• 2026-02-09 — Kieran Klaassen is featured demonstrating the full "How to Make Claude Code Better Every Time You Use It" system alongside Peter Yang. The workflow uses the Compound Engineering plugin for Claude Code CLI with slash commands such as `workflows plan`, `workflows work`, `assess`, and `triage` to run a planning–coding–assessing–codifying loop. The mention also notes that codified learnings are written as Markdown into `/docs/architecture-decisions/` and `/docs/solutions/`, while the root `claude.md` is updated so future planning prompts inherit those rules. Additional details include use of Opus 4.5 and Playwright for end-to-end browser test generation, screen recordings attached to pull requests, and an alias configuration that suppresses interactive permission prompts for unattended sessions.
• 2026-02-10 — Klaassen is mentioned again for demonstrating the Compound Engineering plugin for Claude Code CLI, emphasizing the same iterative loop of planning, coding, assessing, and codifying. The focus remains on capturing insights in a local docs directory so generated code quality improves over time.

Relevance to AI PMs

• Designing repeatable AI delivery systems: Klaassen’s workflow is a practical example of how to turn AI coding into a structured process with explicit stages for planning, execution, review, and knowledge capture.
• Building persistent project memory: The pattern of storing lessons in docs folders and updating `claude.md` shows PMs how to create reusable context that improves future AI outputs without depending on tribal knowledge.
• Operationalizing quality assurance: The use of Playwright-generated end-to-end tests, console inspection, and video artifacts demonstrates how AI workflows can include validation and auditability, not just code generation.

Related

• compound-engineering-plugin — The core plugin Klaassen demonstrates; it enables the structured Claude Code workflow and codification loop.
• claude-code — The execution environment where Klaassen’s slash-command workflow runs.
• peter-yang — Featured alongside Klaassen in the newsletter mention discussing the full system.
• playwright — Used in the demonstrated workflow for automated browser-based end-to-end testing and artifact generation.
• compound-engineering — The broader framework or methodology associated with the plugin and workflow patterns Klaassen showcases.