Give a colleague a skill you made.
At its core, a skill is words: reusable instructions that tell an AI how to do something. A simple skill can travel as a Markdown file—a plain-text document, usually ending in .md—or as the complete text inside an email. Some skills also include scripts, templates or other files, which need to travel with the instructions.
Try it now
Paste it into any AI chat. It has everything your AI needs to guide you through this move.
How to run it
Find out how your organization shares skills. Look for an internal skills library, guidance from your AI team or an established sharing channel. Ask a colleague or the team that supports your AI tools if the route is unclear. Follow your company’s way of sharing when it has one.
Choose the skill you made and give your colleague a short explanation: what it helps with, which AI tool you use it in and an example of when to try it. Share the skill itself, not just a description of it.
For a skill made entirely of text, attach its Markdown file or paste the complete text into an email, preserving the opening name and description and the rest of its formatting. If it refers to bundled scripts, templates or reference files, share the whole skill folder through the approved route. Emailing only the main text would leave those pieces behind.
Many teams use GitHub. GitHub is a website for storing files, tracking changes and working on them together. A repository, often shortened to “repo,” is a project’s shared collection of files. It can hold plain-language instructions as well as software. If your organization uses GitHub for skills, share the link to the skill in the appropriate company repository and check that your colleague can open it. Keep company-only material in its approved internal location.
Have your colleague follow the installation instructions for their AI tool and try the example you supplied. Sending text transfers the skill; their tool may still require an import or installation step before it can recognize the skill. Confirm they can access the files and use it for the intended task.
Background: the Agent Skills format explains what a skill can contain; GitHub’s introduction explains repositories.
How it goes
What the AI will do
- The colleague’s AI uses the shared skill after it is installed or incorporated in the way their tool supports. No prompt is needed to share the files.
What to have ready
- A skill you made, any files it needs and one example of how to use it.
Where you come in
- Check your organization’s sharing method, send the skill and help your colleague try it.
What you'll get
A colleague who can reuse something you taught your AI.
You're done when
Your colleague has the complete skill, can use it in their AI tool and has tried it on a relevant example.
View full AI instructions
Give a colleague a skill you made.
At its core, a skill is words: reusable instructions that tell an AI how to do something. A simple skill can travel as a Markdown file—a plain-text document, usually ending in .md—or as the complete text inside an email. Some skills also include scripts, templates or other files, which need to travel with the instructions.
Step 1 — For the user
Find out how your organization shares skills. Look for an internal skills library, guidance from your AI team or an established sharing channel. Ask a colleague or the team that supports your AI tools if the route is unclear. Follow your company’s way of sharing when it has one.
Step 2 — For the user
Choose the skill you made and give your colleague a short explanation: what it helps with, which AI tool you use it in and an example of when to try it. Share the skill itself, not just a description of it.
Step 3 — For the user
For a skill made entirely of text, attach its Markdown file or paste the complete text into an email, preserving the opening name and description and the rest of its formatting. If it refers to bundled scripts, templates or reference files, share the whole skill folder through the approved route. Emailing only the main text would leave those pieces behind.
Step 4 — For the user
Many teams use GitHub. GitHub is a website for storing files, tracking changes and working on them together. A repository, often shortened to “repo,” is a project’s shared collection of files. It can hold plain-language instructions as well as software. If your organization uses GitHub for skills, share the link to the skill in the appropriate company repository and check that your colleague can open it. Keep company-only material in its approved internal location.
Step 5 — For the user
Have your colleague follow the installation instructions for their AI tool and try the example you supplied. Sending text transfers the skill; their tool may still require an import or installation step before it can recognize the skill. Confirm they can access the files and use it for the intended task.
Step 6 — For the user
Background: the Agent Skills format explains what a skill can contain; GitHub’s introduction explains repositories.
What a skill contains
The instructions and any supporting files that make up a skill.
What GitHub is
GitHub’s introduction to storing files and collaborating.
https://docs.github.com/en/get-started/start-your-journey/what-is-github
Done when
Your colleague has the complete skill, can use it in their AI tool and has tried it on a relevant example.
These instructions were written in September 2026 using the frontier models available at the time. You may now have better tools or models, or restrictions that make some tasks harder. Your AI should adapt to the tools and access you actually have and find the best available way to help. Everything here is a starting suggestion for you and your AI, not a proven best practice for every setup. Feel free to rewrite any of it to better meet your needs.
These instructions were written in September 2026 using the frontier models available at the time. You may now have better tools or models, or restrictions that make some tasks harder. Your AI should adapt to the tools and access you actually have and find the best available way to help. Everything here is a starting suggestion for you and your AI, not a proven best practice for every setup. Feel free to rewrite any of it to better meet your needs.