Learning Center AI
Leadership & Operations Oct 25, 2026 · 6 min By

What's a SKILL.md File, and Why Should Learning Center Leaders Care?

A SKILL.md file lets you write your instructions once and have AI reuse them every time a task comes up. What it is, why it matters for centers, and four skills worth building.

The prompt you keep retyping

If you've used AI for learning center work for more than a few weeks, you probably have a prompt you keep rebuilding. The one that turns visit data into a monthly summary. The one that makes tutor training scenarios sound like real students. Every time, you re-explain your center, your format, and your rules.

A SKILL.md file fixes that. You write the instructions once, save them as a skill, and the AI pulls them in whenever that task comes up.

So what is a SKILL.md file?

It's a plain text file of instructions that teaches an AI tool how to do one specific job your way. Think of it as the training binder you hand a new student worker, instead of re-explaining the job at the start of every shift.

The file has two parts:

  • A short header with a name and a one-sentence description of when to use the skill.
  • The instructions, written in plain language: steps, rules, format, examples of good output.

A skill lives in a folder, and the folder can also hold supporting files like a report template, a rubric, or a style guide. The AI only reads the name and description at first. It opens the full instructions when your request matches, so you can keep a whole shelf of skills without cluttering every chat.

Here's a tiny one:

---
name: center-newsletter-blurb
description: Use when writing a short blurb about a learning center event or service for the campus newsletter or social media.
---

# Center Newsletter Blurb

1. Lead with what the student gets, not what we offer.
2. Keep it under 80 words.
3. Always end with how to book: the scheduling link and our front desk hours.
4. Friendly and direct. No exclamation points, no "don't miss out."

That's it. No code required. The format started at Anthropic and was released as an open standard in December 2025, so the same file can work across several AI tools instead of being locked to one.

Why learning center admins should care

Most of our work is the same handful of tasks done well, over and over, by a staff that turns over every year or two. That's exactly what skills are built for.

  • Consistency across staff. When your coordinator and your grad assistant run the same skill, they get the same format and tone. No more three versions of the monthly report.
  • Institutional memory on paper. Your report structure, your center's voice, your no-show policy wording all live in a file instead of one person's head. When someone leaves, the know-how stays.
  • Less prompt fatigue. You stop pasting the same three paragraphs of context into every chat.
  • Portability. It's plain text. If your campus switches AI tools, your skills come with you, with little or no rewriting.

Four skills worth building for your center

1. Tutor training scenario builder

Training goes better when tutors practice with students who feel real: the one who wants answers, the one who's quietly lost, the one who shows up the night before the exam. This skill generates role-play scenarios on demand.

  • Instructions: the topics your training covers, the student types you want represented, and a fixed output format (situation, what the student says first, what a strong tutor move looks like, three debrief questions).
  • Supporting files: your training topic list or certification checklist, plus one or two scenarios you've already written as examples.
  • Payoff: a fresh, on-target scenario for every training session in about a minute.

2. Monthly usage report writer

You export visit data, then spend an hour turning numbers into something your director or dean will actually read. This skill does the writing part the same way every month.

  • Instructions: the sections you always include (headline numbers, comparison to the same month last year, busiest courses, one notable trend, one-line takeaway), how to round, and what to leave out.
  • Supporting files: last semester's best report as a model.
  • Payoff: a consistent first draft from a de-identified export, ready for your review.

3. Center voice for communications

Faculty outreach emails, social posts, newsletter blurbs, and website updates should sound like the same center wrote them. A voice skill keeps everyone on message.

  • Instructions: tone rules, words you use and avoid, and the details that must always appear (hours, locations, booking link).
  • Supporting files: a short style guide and a few past messages you were happy with.
  • Payoff: student workers can draft outreach that needs light edits, not rewrites.

4. Front desk policy helper

New front desk staff field the same questions all semester: Can I book two sessions a day? What happens after a no-show? Do you cover this course? This skill answers from your handbook, not from the AI's guesses.

  • Instructions: answer only from the attached policies, quote the relevant rule, and say "check with a coordinator" when the handbook doesn't cover it.
  • Supporting files: your staff handbook and course coverage list.
  • Payoff: faster, accurate answers during training and fewer interruptions for you.

Where you can use skills right now

Support depends on which tool your campus licenses. Here's where things stand as of October 2026:

ToolSkill supportHow to add oneClaudeBuilt in, on Free, Pro, Max, Team, and Enterprise plansUpload your skill folder as a ZIP in settings. Team and Enterprise owners can push skills to everyone in the organization.ChatGPTBuilt in for workplace plans, including Edu, since July 2026Ask ChatGPT "Build me a skill..." in a chat, or upload one you wrote.GeminiNative in Google's developer tools (Gemini CLI, Antigravity); not in the regular Gemini app or Gemini EnterpriseWorkaround: paste the SKILL.md instructions into a Gem and select that Gem when you need it.

If your campus is a Gemini school, the Gem workaround gets you most of the benefit. You lose the automatic "pull it in when relevant" behavior, but you keep the reusable, shareable instructions.

Getting started

  1. Pick one task you've prompted three or more times. That's your first skill. Don't start with a giant do-everything skill; small ones you can mix and match work better.
  2. Let the AI draft it. Describe the task, your format, and an example of good output, and ask it to build the skill. Then edit.
  3. Spend extra time on the description line. It's how the AI decides when to use the skill. Name the situations and words that should trigger it.
  4. Keep student data out. Skills hold instructions, not records. De-identify anything you feed into a skill-driven task and follow your campus data policies.
  5. Test, then tighten. Run it on a few real tasks. When the output drifts, add a rule or an example to the file.

The best part is that a good skill is shareable. Build a solid tutor training scenario skill and you can hand it to colleagues at other centers as easily as forwarding a document. That's the kind of thing this field has always been good at.

Sources

AI use statement

Sources come from my own reading, plus articles flagged by Elvis, an AI agent I built in Muse (Meta, 2026). I save the ones worth keeping with a note on why they matter for learning centers. Claude (Anthropic, 2026) helps me review those notes, choose topics, and outline and draft the article and its prompts. I revise every draft, check claims against the sources, and decide what gets published.

I tested the prompts in ChatGPT (OpenAI, 2026), Claude (Anthropic, 2026), Copilot (Microsoft, 2026), Gemini (Google, 2026) and Grok (xAI, 2026).

References

[[shortcode.authors.mark-woolwine]]

← Back to The Feed