agentleFS
DylXBTDylXBT/claude-orchestrator-skills/skill-author/SKILL.md

skill-author

Skill114 starsChanged 5 months agoMITView on GitHub

Package a task the user repeats into a reusable SKILL.md skill folder, so it runs the same every time instead of depending on memory. Use when someone re-explains the same task to an AI assistant on a schedule. Meta-skill — works for any recurring task in any field.

23 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

2 rules. 0 are common, 1 is distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 18 edited versions we read keep these.

  • 1

    Distinctive

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Write the description carefully — it's what an agent reads to decide *when* to use the skill, so name the task and the trigger plainly.

    line 51
  2. Distinctive:

    Run the new skill against a real brief.

    line 55

“Common” is measured: kept by at least 6 in 10 of the 18 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Skill Author

If you re-explain the same task every week, you're paying for it in time and in quality that swings with your memory. A skill packages the task once: when to run it, how the output gets made, what "done" looks like, and a couple of examples. This skill helps the user write one.

Role-agnostic: a recurring task is a recurring task, whether it's a weekly report, a candidate-screening summary, a design QA pass, or a release note.

Step 1 — Interview

Ask the user:

  1. What task do you repeat? Something they make on a cadence. If they name several, pick the highest-friction one.
  2. When should it run, and what does it need? The trigger and the inputs.
  3. How is it made? The steps, in order. Pull these out of them — they usually know but have never written it down.
  4. What does "done" look like? Sections, length, format, structure.
  5. Got examples? One or two of their best past versions to anchor quality.

Step 2 — Write the SKILL.md

Produce a standard Agent Skill folder: <skill-name>/SKILL.md with this shape:

---
name: <kebab-case-name>
description: <what it does + when to use it — this is how the agent decides to load it>
---

# <Title>

<one-line purpose>

## Trigger
<when to run · required inputs>

## Steps
1. <step>
2. <step>
...

## Output format
<what "done" is: sections, length, structure>

## Examples
<1–2 references, or links to example files>

Write the description carefully — it's what an agent reads to decide when to use the skill, so name the task and the trigger plainly.

Step 3 — Dry-run

Run the new skill against a real brief. If the output drifts, the fix is usually a sharper step or a clearer "done" definition. Tighten, then save.

Step 4 — Hand-off

Tell the user where it lives (~/.claude/skills/<name>/ for Claude Code, or their agent's skills dir), and how to share it — a public GitHub repo is installable by others with npx skills add <owner>/<repo>.

Rules

  • The description field earns its keep — vague descriptions never get loaded.
  • Number the steps so output is identical whether the user is calm or rushed.
  • Anchor on real examples, not adjectives.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.