Teach it the patterns you repeat.

Teach it once. It knows it forever.

A skill is a named recipe — instructions, templates and examples written in plain language — that your agents reach for when the task fits. Each skill has a trigger (when it activates), a process (what it does, step by step), and an output shape (what it produces). All written in English, not code.

Skills live in a shared library. You grant specific skills to specific agents. Editing a skill once updates it for every agent that uses it. If an agent needs a specialised version, you fork the skill into a copy and tailor it.

The same way you build everything else in 1Presence — through conversation. You describe the pattern in your own words. Presence shapes it into something precise:

• You: "Whenever I say 'prep me', check my next meeting, pull context from my vault, write a one-page brief."
• Presence: "Got it. I'll trigger on 'prep me' or 'brief me for X'. I'll pull the next event from your primary calendar, search vault Projects/ and People/ for context, and produce a brief under Daily/Prep/ with sections: Attendees, Context, Goals, Where we left it. Want me to also scan email threads with the attendees?"
• You: "Yes, the last couple of weeks."
• Presence: "Saved. The skill is live."

Total time: under a minute. The result is a precisely-worded skill that triggers reliably, structures its output, and can be improved over time by saying things like "actually, also include a draft opener I could use."

The interesting thing about skills is that they accumulate. A user who has been on the product for six months has thirty skills — each saving them a few minutes of explanation, each producing a consistent output, each editable in plain language. Skills become a small private operating system for the way you work.

And because skills live in a shared library rather than buried in code, they are yours: written in plain language, browsable on each skill's own page, editable any time, and portable between your agents. Open a skill to see exactly what it does, which agents use it, and when it last ran.

Skills are the building blocks of agents and workflows. An agent is essentially an identity, a set of skills, and the connectors it can reach. A workflow is a sequence of stages, each one run by an agent drawing on its skills. Once you have built up your skill library, composing new workflows becomes a matter of arrangement, not invention.