Skill Extraction
A skill is the crystallized residue of a workflow you already
proved by hand — so you build the thing first and package it
last. The rule of thumb: you don't start with a skill, you
end with a skill. Have an idea, get the agent to build a first
version, and give it round after round of feedback until the output hits
your standard and taste; only then tell the agent "take the workflow we
just did and turn it into a skill," and separately design the onboarding
(what context the skill must ask a user for). Packaging cold — writing
SKILL.md before the workflow works — inverts the
dependency: you'd be freezing a process you haven't validated. This is
why the move is accessible to non-technical authors: the hard part is
doing the task well and noticing your own taste, which is human work;
the freeze is a single instruction to the agent.
Skill extraction is the provenance rule the authoring cluster leaves implicit. Skill-Driven Loop Development insists you pre-validate skills before looping over them and Skill Authoring Checklist governs a single skill's internals — but neither says where a skill comes from. The answer here: from a proven, hand-run workflow, extracted at the end.
Claims
- You don't start with a skill, you end with a skill — build and battle-test the workflow by hand, then have the agent package it. principle — durable: a skill encodes how you want a task done, which you can only know after doing it well; packaging before the workflow is proven freezes an unvalidated process. Mirrors the "already validated by hand" bar in Skill-Driven Loop Development.
- Extraction has two parts: freeze the proven workflow, then design the onboarding — figure out what context the skill must elicit from a user so the agent knows what to ask for. best practice — context: turning a personal workflow into something others run; the workflow alone is ~70% of a skill, the missing piece is the user journey (see the onboarding flow in Stop Making PowerPoints: Vibe-Coding HTML Slides as a Skill).
- Anyone can extract a skill, especially non-technical people — a skill is your expertise, taste, best practices, and workflow frozen into something an agent can run. observation — the video's definition; the leverage is that the valuable input (taste/judgment) is human, and the technical packaging is delegated to the agent.
Related
- Skill-Driven Loop Development — the sibling rule: pre-validate skills before looping. Same "battle-tested first" instinct; that page is about composition (validated skills as loop steps), this one about origination (where a skill comes from).
- Skill Authoring Checklist — governs the internals of the skill you extract (trigger, structure, steering, pruning); skill-extraction is the step before the checklist applies.
- Vibe Coding — the build phase skill-extraction sits on top of: rounds of prompt→generate→inspect feedback are exactly how the workflow gets proven before it's frozen.
- Reusable Workflow Library — where an extracted skill goes to be catalogued and reused; "a workflow that worked once is a durable artifact" is the same instinct one step later.
- Knowledge Work as Code — why the payoff accrues to non-programmers: the workflow being frozen is ordinary knowledge work, and the freeze turns taste into a shippable artifact.
- Distillate: Stop Making PowerPoints: Vibe-Coding HTML Slides as a Skill