skill-factory
Skill0 starsChanged 5 months agoMITView on GitHub
Turn any workflow, conversation, or demonstrated process into a reusable SKILL.md. Use when the user says "turn this into a skill", "make a skill from this", "save this workflow", "remember how to do this", or after completing a complex multi-step task that should be repeatable. Also triggers on "tanítsd meg magad", "jegyezd meg ezt a folyamatot", "csinálj ebből skill-t".
13 other projects use a version of this, 1 of them edited.
Adopted by
Most recently taken up by
- Szotasz/marveenedited, first seen Nov 7, 2026
- hamgarian/loot-skilltheir own version, first seen Oct 25, 2026
- kannaka-labs/Kannaktopustheir own version, first seen Oct 13, 2026
- superXdev/skillstheir own version, first seen Sep 15, 2026
- akijain2000/skill-factorytheir own version, first seen Sep 5, 2026
- nyldn/claude-octopustheir own version, first seen Aug 13, 2026
Other versions13 projects use a version of this.
Word for word · 0
None found.
Edited · 1
- Szotasz/marveen★ 13156% same
Wrote their own version · 12
- nyldn/claude-octopus★ 4.4ksame name
- ITSpecialist111/Copilot-Cowork-Skills★ 48same name
- immacualate/claude-forge★ 5same name
- akijain2000/skill-factory★ 4same name
- bonez-io/re_gent★ 4same name
- echoleesong/claude-skills-plugin★ 4same name
- shivamd1810/claude-skill-factory★ 4same name
- hamgarian/loot-skill★ 2same name
- kingjly/skill-factory★ 2same name
- lambda610/skills★ 2same name
2 more not listed
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository2 other files it gives its agents.
- gemini-brainstormSkill
- share-clean-packageSkill
The file
We split this skill into the rules it gives an agent
10 rules. 0 are common, 8 are distinctive, and 2 are very specific to this project.
0
Common
Most of the 12 edited versions we read keep these.
8
Distinctive
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 14
After a complex task (5+ tool calls) that could be reusable
- Distinctive: line 15
When user corrects your approach and the correction is generalizable
- Distinctive: line 24
Trigger conditions: When should this skill activate?
- Distinctive: line 34
Don't just copy the specific instance.
- Distinctive: line 82
If the workflow involves scripts or templates:
- Distinctive: line 95
Test the skill mentally:
- Distinctive: line 103
Overfitting to one example: Don't just save the exact steps you did.
- Distinctive: line 105
Missing error handling: If you hit errors during the original workflow, document them in Pitfalls.
- Very specific: line 106
Too long: Keep SKILL.md under 500 lines.
- Very specific: line 107
Duplicate skills: Before creating, check your existing ~/.claude/skills/ for similar skills.
“Common” is measured: kept by at least 6 in 10 of the 12 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.
Files
Skill Factory
Convert demonstrated workflows into reusable skills. This is a meta-skill: it creates other skills.
When to Use
- User explicitly asks to create a skill from a workflow
- User says "remember how to do this" or "save this process"
- After a complex task (5+ tool calls) that could be reusable
- When user corrects your approach and the correction is generalizable
- Hungarian triggers: "tanítsd meg magad", "csinálj skill-t", "jegyezd meg"
Procedure
Step 1: Extract the Workflow
Analyze the current conversation (or specified workflow) and identify:
- Trigger conditions: When should this skill activate?
- Input: What does the skill need to start?
- Steps: What are the concrete steps (in order)?
- Tools used: Which tools/commands were involved?
- Decision points: Where did you need to make choices?
- Error handling: What went wrong and how was it fixed?
- Output: What's the expected result?
Step 2: Generalize
Don't just copy the specific instance. Abstract it:
- Replace specific file names with
[input-file]patterns - Replace specific URLs with
[target-url]patterns - Identify which parts are constant vs. variable
- Note any prerequisites or dependencies
- Think about edge cases the original workflow didn't cover
Step 3: Write SKILL.md
SKILL_NAME="[kebab-case-name]"
mkdir -p ~/.claude/skills/$SKILL_NAME
cat > ~/.claude/skills/$SKILL_NAME/SKILL.md << 'EOF'
---
name: [skill-name]
description: [What it does + when to trigger. Be specific and "pushy" -- include multiple trigger phrases so the skill activates when needed.]
---
# [Skill Name]
## When to Use
[List concrete trigger conditions and contexts]
## Prerequisites
[Dependencies, tools, access needed -- skip if none]
## Procedure
1. [First step -- be specific, include commands]
2. [Second step]
...
## Pitfalls
- **[Problem]**: [How to solve it]
## Verification
- [How to confirm the result is correct]
## Examples
**Example 1:**
Input: [what the user said]
Output: [what was produced]
EOFStep 4: Add Supporting Files (if needed)
If the workflow involves scripts or templates:
mkdir -p ~/.claude/skills/$SKILL_NAME/scripts
mkdir -p ~/.claude/skills/$SKILL_NAME/referencesscripts/: Executable code for deterministic/repetitive tasksreferences/: Documentation loaded into context as neededassets/: Templates, icons, or other static files
Step 5: Validate
Test the skill mentally:
- Would the description trigger on a realistic user message?
- Are the steps clear enough to follow without the original context?
- Are edge cases covered in Pitfalls?
- Is the SKILL.md under 500 lines?
Pitfalls
- Overfitting to one example: Don't just save the exact steps you did. Generalize so it works for similar but different inputs.
- Too vague descriptions: The description field is the primary trigger. Be specific and include multiple phrasings.
- Missing error handling: If you hit errors during the original workflow, document them in Pitfalls.
- Too long: Keep SKILL.md under 500 lines. Move large content to
references/subdirectory. - Duplicate skills: Before creating, check your existing
~/.claude/skills/for similar skills. Patch instead of creating a new one.
Skill Quality Checklist
Before finalizing, verify:
- Description includes multiple trigger phrases
- Steps are numbered and concrete
- Commands are copy-pasteable (no pseudocode)
- Variables are clearly marked with
[brackets] - Pitfalls section has at least one entry
- Verification section explains how to confirm success
- Under 500 lines
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
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.

