Create skill from workflow
Skill az9713/youtube-transcript-to-skills/.claude/skills/create-skill-from-workflow
Extract Claude Code skills from AI workflow YouTube transcripts. 10 ready-to-use skills + a meta-skill that automates the extraction process. Built from John Kim's 50 Claude Code tips.
npx -y skills add az9713/youtube-transcript-to-skills --skill create-skill-from-workflowAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
2 things to look at
- no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
- 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
What its author says it does
Copied from the file, not written here
Turn what we just did into a reusable skill. Captures the current workflow as a new SKILL.md file.
SKILL.md
2.7 KB, 626 tokens by cl100k_base, as published. Nobody here has run it
Create Skill from Workflow
Turn a workflow you just performed into a reusable SKILL.md file.
Usage
/create-skill-from-workflow <skill-name>
Procedure
Step 1: Identify the Workflow
Review the current conversation to identify:
- What task was just completed?
- What steps were taken (in order)?
- What decisions were made along the way?
- What tools/commands were used?
- Were there any gotchas or mistakes that should be avoided?
Ask the user if anything is unclear about the workflow scope.
Step 2: Determine Skill Properties
Ask or infer:
| Property | Question |
|---|---|
| Name | Use $ARGUMENTS as the skill name (kebab-case) |
| Description | One sentence: what does this skill do and when should it be used? |
| Model-invocable? | Should Claude auto-invoke this, or only when user asks? Default: user-only (disable-model-invocation: true) |
| Forked? | Should this run in an isolated context? Default: no (runs inline) |
| Arguments? | Does the skill need input? If so, what? |
Step 3: Extract the Steps
Convert the observed workflow into a structured procedure:
- Remove one-off decisions specific to this instance
- Generalize file paths and names into patterns
- Add decision points where the workflow could branch
- Include verification steps ("confirm X before proceeding")
- Add error handling ("if X fails, try Y")
Step 4: Write the SKILL.md
Create the file at .claude/skills/<skill-name>/SKILL.md:
---
name: [skill-name]
description: [one-sentence description]
disable-model-invocation: [true/false]
argument-hint: [argument description, if needed]
---
# [Skill Title]
[Brief description of what this skill does and when to use it.]
## Usage
`/[skill-name] [arguments if any]`
## Procedure
### Step 1: [Name]
[Instructions...]
### Step 2: [Name]
[Instructions...]
[...continue for all steps]
Step 5: Verify the Skill
- Check that the file exists at the correct path
- Confirm the skill name doesn't conflict with existing skills
- Tell the user: "The skill is now available as
/[skill-name]. Try invoking it to test."
Tips for Good Skills
- Be specific: "Run
npm test" is better than "run tests" - Include the WHY: Explain why each step matters
- Add escape hatches: What should the user do if a step fails?
- Keep it focused: One skill = one workflow. Don't combine unrelated tasks.
- Use $ARGUMENTS: If the skill needs input, reference
$ARGUMENTSin the instructions