agentleFS
BrandeisPatrickBrandeisPatrick/silo-harness/skills/create-skill/SKILL.md

create-skill

Skill2 starsChanged 5 days agoMITView on GitHub

Writes a new skill or improves an existing one, as a SKILL.md with frontmatter and step-by-step instructions, then loads it to check it. Use when the user asks to make, write, save or fix a skill, turn a procedure into a skill or slash command, or says "remember how to do this".

148 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use review-bugbot, migrate-to-skills and 10 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • 0mlkrizzz655335vSkill

    review-bugbot

    缺陷导向代码审查。用 Bugbot 子代理找 bug。

    Also in 5 projects that use this

  • 32103210Skill

    migrate-to-skills

    Convert 'Applied intelligently' Cursor rules (.cursor/rules/*.mdc) and slash commands (.cursor/commands/*.md) to Agent Skills format (.cursor/skills/). Use when you want to migrate rules or commands to skills, convert .mdc rules to SKILL.md format, or consolidate commands into the skills directory.

    Also in 8 projects that use this

  • 32103210Skill

    update-cursor-settings

    Modify Cursor/VSCode user settings in settings.json. Use when you want to change editor settings, preferences, configuration, themes, font size, tab size, format on save, auto save, keybindings, or any settings.json values.

    Also in 8 projects that use this

  • PandelisZSkill

    update-cli-config

    View and modify Cursor CLI configuration settings in ~/.cursor/cli-config.json. Use when the user wants to change CLI settings, configure permissions, switch approval mode, enable vim mode, toggle display options, configure sandbox, or manage any CLI preferences.

    Also in 7 projects that use this

  • yawningwhaleAISkill

    build-premium-website

    Build a premium, animated marketing website (React + Vite + Tailwind + GSAP) for any industry. Use when the user asks to "build a website", "make a landing page for [business]", "build a marketing site", or wants a high-end animated company website. Adapts color, copy, services, and signature animation to the industry.

    Also in 6 projects that use this

  • yawningwhaleAISkill

    instantly-campaign

    Create a new Instantly cold-email campaign in the proven 5-email structure of the user's top-performing campaign (~16% reply rate). Use when the user wants to launch a new outbound campaign, write a cold email sequence, create an Instantly campaign, draft outreach for a new service or niche, or set up a sales sequence.

    Also in 5 projects that use this

  • +1 more from 32103210

The file

We split this skill into the rules it gives an agent

11 rules. 0 are common, 5 are distinctive, and 6 are very specific to this project.

  • 0

    Common

    Most of the 138 edited versions we read keep these.

  • 5

    Distinctive

    This author's own choices. Worth a look.

  • 6

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Keep it to one job. If the request holds two, write two skills.

    line 11
  2. Very specific:

    Default: the project, <project>/.silo-harness/skills/<name>/, where <project> is the session folder (or its git root when the skill is for …

    line 15
  3. Very specific:

    Add argument-hint when the skill takes input.

    line 41
  4. Very specific:

    Add disable-model-invocation: true only when the skill has side effects the user should start themselves (deploys, sending messages).

    line 42
  5. Distinctive:

    No filler, no greetings, no restating the description.

    line 44
  6. Very specific:

    Keep SKILL.md under about 200 lines.

    line 45
  7. Very specific:

    Create the folder and write SKILL.md and any supporting files.

    line 48
  8. Distinctive:

    Read the result: the body must come back whole, the base directory must be the folder you chose, and the files list must show every …

    line 53
  9. Distinctive:

    If it lists problems, fix each one and load it again.

    line 54
  10. Very specific:

    If the tool says the skill is run by the user only, that is the disable-model-invocation you set.

    line 55
  11. Distinctive:

    Read the description once more as the model would: would a user's request for this job match it, and would unrelated requests not?

    line 56

“Common” is measured: kept by at least 6 in 10 of the 138 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

A skill is a folder <name>/SKILL.md: YAML frontmatter, then instructions in Markdown, plus any files beside it. format.md in this skill's folder has the exact rules and every problem message; read it when a rule below is not enough.

1. Find out what the skill is for

  • Take the job from the user's words, the conversation, or $ARGUMENTS. If the procedure was just done in this session, use the steps that actually worked.
  • Settle two things before writing: what the skill does, and when it should trigger (the phrases a user would say). Ask the user only if you cannot infer either one.
  • Keep it to one job. If the request holds two, write two skills.
  • To change an existing skill, load it with the skill tool first; the result gives its folder.

2. Choose where it lives

  • Default: the project, <project>/.silo-harness/skills/<name>/, where <project> is the session folder (or its git root when the skill is for the whole repository).
  • The user's own, ~/.silo-harness/skills/<name>/, when it is personal or useful across projects.
  • Tell the user which one you chose and why, in one sentence.
  • A project skill of the same name hides the user's and the built-in one; pick another name unless replacing it is the point.

3. Pick the name

  • Lower-case letters, digits and single hyphens, at most 64 characters: deploy-staging, release-notes.
  • A verb or a clear noun phrase. The folder name and name: must be the same.
  • The user will type it as /name, so keep it short.

4. Write the description

One or two sentences, at most 1,024 characters, in the third person: what it does, then "Use when …" with the words a user would actually say. The model picks skills from this line alone.

  • Good: Drafts release notes from the commits since the last tag, grouped into New, Improved and Fixed. Use when the user asks for release notes, a changelog entry, or what changed since the last release.
  • Bad: Release notes helper. (no trigger, no outcome)

5. Write SKILL.md

---
name: release-notes
description: Drafts release notes … Use when the user asks for …
argument-hint: <version>
---
1. Find the last tag with `git describe --tags --abbrev=0`.
2. …
  • Add argument-hint when the skill takes input; use $ARGUMENTS, $1, $2 in the body where it goes.
  • Add disable-model-invocation: true only when the skill has side effects the user should start themselves (deploys, sending messages). Add user-invocable: false only for background knowledge nobody types.
  • The body: numbered steps in the imperative, in working order, each one something you can check was done. Name exact commands, files and tools. Say what "done" looks like.
  • No filler, no greetings, no restating the description. No secrets, keys, tokens or personal paths; read them from the environment or ask.
  • Keep SKILL.md under about 200 lines. Put long reference material, templates and scripts in files beside it and refer to them by relative path (see checklist.md, run scripts/check.sh). Use ${SKILL_DIR} when a command needs the absolute folder.

6. Write the files

  • Create the folder and write SKILL.md and any supporting files.
  • Make scripts executable (chmod +x) and run each once to see it works.

7. Verify

  1. Load it: call the skill tool with the name. It re-reads the skill folders, so a new skill is found.
  2. Read the result: the body must come back whole, the base directory must be the folder you chose, and the files list must show every supporting file.
  3. If it lists problems, fix each one and load it again. Repeat until there are none.
  4. If the tool says the skill is run by the user only, that is the disable-model-invocation you set; it still lists any problems.
  5. Read the description once more as the model would: would a user's request for this job match it, and would unrelated requests not?

8. Finish

Tell the user, briefly:

  • the skill's name and folder;
  • how to run it: /name (with the argument hint, if any), or just ask for the job in their own words;
  • that Settings → Skills lists it.

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.