Turn Any Process Into a 2-Minute Skill

How it works

Turn Any Process Into a 2-Minute Skill, step by step.

1. 01
   The Problem

   You keep re-explaining the same process

   You've figured out how to debrief a call. How to write a follow-up. How to draft a proposal. You know the steps, the format, the edge cases.

   But every time you ask AI to do it, you re-explain. The prompt you wrote last week is buried in a chat history. The instructions you tweaked are gone. You're training AI from scratch, every time.

   Same process, re-explained from scratch

2. 02
   The Fix

   Skills, not prompts

   A prompt is a one-shot instruction. A skill is a reusable capability. It has a name, trigger conditions, step-by-step process, push-back rules, output format, and quality gates.

   Save it once as a markdown file. Invoke it with /skill-name. It reads the right files, follows the right steps, challenges you when something's off, and outputs in the right format. Every time.

   Prompt = one-shot
   Skill = reusable capability

3. 03
   The Anatomy

   Seven parts, one file

   Every skill has the same anatomy: frontmatter (triggers), context (what to read first), process (step-by-step), push-back rules (when to challenge), output format (what the result looks like), quality gates (self-check), and examples (calibration).

   The push-back rules are the moat. Anyone can write 'summarize this call.' A skill knows when to say 'you had a risk flagged for this person. Did it come up?'

   FrontmatterName, triggers, description

   ContextWhat to read before starting

   ProcessStep-by-step execution

   Push-Back RulesWhen to challenge the user

   Output FormatWhat the result looks like

   Quality GatesSelf-check before finalizing

   ExamplesShow, don't tell
4. 04
   The Chain

   Skills compound

   A debrief skill feeds an entity-update skill. The entity update feeds your morning brief. The morning brief surfaces a follow-up. The follow-up triggers a meeting prep skill.

   You're not building a collection of prompts. You're building a chain where each skill makes the next one smarter. That's why the context section matters: skills read each other's output.

   Debrief → Entity update → Morning brief → Follow-up

5. 05
   Getting Started

   Pick your most repeated process

   Start with the thing you do every week that takes the same shape. Meeting prep. Weekly reports. Client follow-ups. The process you could teach someone in 5 minutes but AI still gets wrong.

   Write the seven parts. Start rough. The skill gets better every time you use it and notice what's missing. First version is never the final version. That's the point.

   Write once → invoke forever → iterate as you go

6. 06

   The guide that teaches you the anatomy.

   Inside: the 7-part skill anatomy with a full worked example (a call debrief skill from frontmatter to quality gates). Plus the design principles that make skills compound instead of collide.

   Drop your email below. I'll send you the guide plus the walkthrough video.