How To Prompt Like An Anthropic Engineer
Short summary
The author argues that treating prompts as disposable is wasteful—instead, build modular "skills" (organized folders with instructions and scripts) that you improve over time and chain together. He breaks down 17 tactics from Anthropic engineers into 6 layers, showing how to structure workflows so improvements compound across tasks rather than starting from scratch each time.
- The author rebuilt a 600-line SEO prompt into six modular skill files (research, audit, metadata, internal links, GSC monitoring, title scoring), and improving one now improves every workflow that uses it.
- The core shift is from one-off prompting to building reusable "skills"—folders with a SKILL.md file, scripts, and references—that Claude runs identically every time and that get refined after each failure.
- Anthropic engineers' 17 tactics get grouped into 6 layers, with the top priority being: stop writing custom prompts for repetitive tasks and instead build slash-command-invoked skills.
- The lasting value comes from the infrastructure built on top of Claude, not from any single prompt, which evaporates once the session ends.
Full summary
The core argument is that most people approach Claude wrong. They treat it like a search engine—write a prompt, get an answer, move on. Anthropic's engineers work differently. They build reusable systems: organized folders with focused instructions and scripts, connected through slash commands and agents. The author watched talks from Anthropic engineers at a May 2026 event and realized his entire SEO workflow—a sprawling 600-line document crammed into one file—was backwards. He split it into six focused files (research, audit, metadata, internal links, GSC monitoring, title scoring), wired them together, and now when he improves one skill, every workflow using it automatically gets better. That's the compounding advantage Anthropic's people keep talking about.
The shift is from prompting to system design. When you write a one-off prompt, you get a one-off result. The value evaporates when the session ends. When you build a skill—a folder with a SKILL.md file, scripts, and references—Claude runs it the same way every time, and you improve it after each failure. Six months in, it's doing work your day-one self couldn't have even imagined. The author groups Anthropic's tactics into six layers, with the first being the most important: stop writing custom prompts for repetitive tasks. Build skills instead. Invoke them with a slash command. Run the same task the exact same way every single time. This is where your time actually accumulates into something that matters, not in the prompt itself but in the infrastructure you build on top of what Anthropic provides.