Convertly Tools

SKILL.md Generator

Describe a task you keep asking Claude to do and get a SKILL.md draft in the published format: frontmatter, instructions, examples. A model writes it, this page checks it against the format rules, and you edit it before you save it.

Lowercase words joined by hyphens. Anything else is tidied for you.

Optional: rules, examples, and output format

Before you generate: what you type here is sent to Anthropic’s API to write the draft. It is the one tool on this site that does not run in your browser. Convertly Tools does not store it. Do not include secrets or private data in the description. How this is handled.

How your text is handled: This is the one tool on the site that does not run on your device. The text in the form is sent to Anthropic's API through our server to write the draft, and the reply comes back to your browser. Convertly Tools does not store or log it. Read the privacy note.

How to generate a SKILL.md

  1. Name the skill, say what it should do, and say when Claude should reach for it. The optional fields take rules, examples, and the output shape you want.

  2. Press Generate SKILL.md. The description is sent to Anthropic's API and a draft comes back, usually in ten to thirty seconds.

  3. Read the draft in the editor. Every edit is re-checked against the frontmatter rules underneath it.

  4. Copy the Markdown or download it as SKILL.md, then save it in a folder named after the skill.

What a SKILL.md is

A skill is a folder containing a SKILL.md file: YAML frontmatter with a name and a description, then Markdown instructions. Claude loads every skill's name and description at startup and reads the full file only when a request matches, which is why the description carries so much weight. Claude Code reads skills from .claude/skills/ in a project or ~/.claude/skills/ for your account; claude.ai takes them as a zip upload; the API takes them through the Skills endpoints.

The format is published as the open Agent Skills specification, and Anthropic's own authoring guide adds conventions on top: keep the body concise, prefer specific steps over general advice, give examples where output quality depends on them, and stay under about 500 lines. This tool follows both when it writes, and checks the frontmatter rules when you edit.

What this tool sends, and where

This is the one tool on Convertly Tools that does not run in your browser. Writing a good skill needs a language model, and the model runs on Anthropic's servers. When you press Generate, the text in the form is posted to a small function on our server, which forwards it to Anthropic's Messages API with our key and returns the reply to your browser.

Convertly Tools does not store the description or the draft, does not log their contents, and does not send them to analytics. Anthropic handles the request under its own API data policy. The practical rule is the same as for any AI service: describe the task, but keep credentials, customer data, and anything confidential out of the description. If your skill needs a secret, write a placeholder and fill it in locally.

What the checks cover, and what they don't

After every edit the page checks the file against the published rules: the frontmatter must open on line one; name is required, 1 to 64 characters, lowercase letters, numbers and single hyphens, and must not contain the words "anthropic" or "claude"; description is required, under 1024 characters, and free of XML tags; the optional compatibility field is capped at 500 characters; metadata must be a simple map. It also flags Claude Code-only fields, which stop a file uploading to claude.ai or the API, and a few authoring slips such as claiming the skill has been tested.

A passing file has valid structure. It says nothing about whether the instructions are good, whether Claude will trigger the skill when you expect, or whether the steps suit your project. Those come from reading the draft, correcting anything the model assumed, and trying the skill on a real task. Anthropic's guide recommends writing a few test scenarios before relying on a skill, and that advice applies here.

When the draft makes assumptions

The model is told not to invent tools, file paths, commands, or facts about your project. When the form leaves something essential unsaid, the draft should add a short Assumptions section naming what it guessed, or use a bracketed placeholder like [path to the config file]. Treat those as the first things to correct. A skill that names a script that does not exist is worse than one that asks.

Good to know

  • Save the file as SKILL.md inside a folder with the same name as the name field. The folder name must match.

  • The draft is a starting point. Read all of it before Claude does.

  • Fields such as disable-model-invocation, model, and paths are Claude Code extensions. They are fine in a Claude Code skill and cause an upload to claude.ai or the API to fail.

  • Six generations per ten minutes from one connection. Editing the draft in the box is free and unlimited.

  • Your description goes to Anthropic's API. Nothing you type is stored by Convertly Tools.

  • New to skills? How to create a Claude skill covers the whole process, and the SKILL.md format explained lists every field this page checks.

  • Tidying JSON that a skill produces or consumes? The JSON Formatter runs entirely in your browser.

Frequently asked questions

Is my description uploaded anywhere?

Yes. This is the one tool on the site that sends what you type off your device: to our server, and from there to Anthropic's API, which writes the draft. Convertly Tools does not store or log it. Keep secrets and private data out of the description.

Does passing the checks mean Claude will use the skill correctly?

No. The checks confirm the file has valid frontmatter and a sensible shape. Whether Claude triggers it at the right moment and follows it well depends on the description and the instructions, which only reading and trying the skill can tell you.

Where do I put the file?

For Claude Code, in a folder under .claude/skills/ inside your project, or ~/.claude/skills/ for all projects, named after the skill: for example .claude/skills/writing-release-notes/SKILL.md. For claude.ai, zip the folder and upload it under Settings. For the API, upload it through the Skills endpoints.

Why does the name have to look like that?

The specification requires 1 to 64 characters of lowercase letters, numbers, and hyphens, with no leading, trailing, or doubled hyphens, and Anthropic additionally reserves the words "anthropic" and "claude". The folder must carry the same name. Whatever you type is tidied into that form before the draft is written.

Can I use the same file in Claude Code and claude.ai?

Yes, as long as the frontmatter only uses the fields in the specification: name, description, license, compatibility, metadata, and allowed-tools. Claude Code accepts extra fields such as model or paths, but a file containing them is rejected when uploaded to claude.ai or the API. The checker warns you when one is present.

Which model writes the draft?

A current Claude model through Anthropic's API. The page does not promise a specific one, because the choice is a server setting that can change. Whichever it is, the draft is still a draft and needs your review.

Guides for this tool

  • How to Create a Claude SkillA Claude skill is a folder holding a SKILL.md file that teaches Claude a task you keep repeating. To create a Claude skill you need no code: a folder, a short file with two required fields, and a test. This guide covers the five steps, with the rules from the published format, and points to the deeper guides where a step needs one.
  • How to Write a SKILL.md FileWriting a SKILL.md file is mostly a writing problem, not a formatting one. The format is two required fields and some Markdown. What decides whether Claude uses the skill is the description, and what decides whether the result is any good is how the instructions are written. This guide covers both, using the rules in Anthropic's authoring guide and the Agent Skills specification.
  • Claude SKILL.md Format ExplainedThe SKILL.md format has two parts: YAML frontmatter that tells Claude when to use the skill, and a Markdown body that tells it how. Two fields are required, four more are in the open specification, and Claude Code understands a further set that no other surface accepts. This page lists every field with its limits, as published in the Agent Skills specification and Anthropic's documentation.
  • How to Build a Claude Code SkillA Claude Code skill is a SKILL.md file in a folder that Claude Code watches. That much is shared with every other surface. What is specific to Claude Code is where it looks, how a skill becomes a slash command with arguments, and the frontmatter that controls who can invoke it and which tools it may use without asking. This guide covers those parts, as documented in the Claude Code skills reference.
  • Claude Skills vs MCPClaude Skills and MCP get compared because both extend what Claude can do, but they extend different things. A skill changes what Claude knows: it is a folder of instructions loaded when a task matches. MCP, the Model Context Protocol, changes what Claude can reach: it is an open standard for connecting to outside tools and data through servers. Most real setups use both, and this guide explains where the line falls.

Related tools

SKILL.md Generator is one of 2 developer tools, format and check data files, and draft skills for Claude.