A skill is a packaged workflow attached to a colleague. Once a skill is in place, the colleague runs it the same way every time — the steps, the checks, the output shape, all consistent. You can add skills three ways.Documentation Index
Fetch the complete documentation index at: https://docs.eluu.ai/llms.txt
Use this file to discover all available pages before exploring further.
From a GitHub URL
If you’ve got a skill repo or a teammate has shared one, paste the GitHub URL into the colleague’s skills section. Eluu pulls it in, parses it, and attaches it to the colleague. A common pattern is teams maintaining a private GitHub repo of their company’s skills —acme-skills/morning-brief, acme-skills/customer-health-check. Pull whichever one you need into whichever colleague should have it.
From your team library
You can promote a skill to your team’s shared library. Once it’s there, anyone building a colleague can check the box to attach it. Useful for company-standard workflows.Have the colleague add it themselves
You don’t have to manage skills by hand. If you’re working with a colleague and the right move is to add a skill, just ask:Find a meeting transcription skill on GitHub and add it to yourself.
Look for a research skill that’s good for competitive analysis. Add it to your library.The colleague searches, evaluates, and pulls the skill in. You’ll see what they imported. The skill is available immediately.
How a colleague uses a skill
Once attached, a skill is available in two ways:- Implicitly — when you describe a task that matches a skill’s purpose, the colleague runs the skill without you naming it.
- Explicitly — type
/in chat and pick the skill by name.
pipeline-review is going to do a much better job than a generic ask, because the skill captures the steps you’ve decided are right.
Improving a skill
Skills aren’t frozen. When one produces something off, tell the colleague:The brief was too long. Cap it at 600 words.
When the customer name has multiple words, only use the first word as the email greeting.The colleague updates the skill so the same mistake doesn’t repeat. Over time, your skills get tighter and more reliable. See skill creation for the technique behind making skills robust.
Where to next
Skill creation in depth
Make skills that handle edge cases.
Verification loops
Build trust into every skill run.