by Helpercraft
Craft your own AI agent: a real Agent Skill (SKILL.md) with a job, rules and a personality. Keep your team in one folder; your AI suggests who fits and asks first. Free, private, one file.
# Add to your Claude Code skills
git clone https://github.com/Helpercraft/helpercraftSee how helpercraft compares with popular alternatives.
helpercraft is an open-source ai agents skill for AI coding assistants such as Claude Code, Codex CLI, and ChatGPT, built by Helpercraft. Craft your own AI agent: a real Agent Skill (SKILL.md) with a job, rules and a personality. Keep your team in one folder; your AI suggests who fits and asks first. Free, private, one file. It has 50 GitHub stars.
helpercraft's catalog security scan is still queued. You can run an instant dependency and prompt-injection check now with the "Scan for vulnerabilities" button above.
Clone the repository with "git clone https://github.com/Helpercraft/helpercraft" and add it to your Claude Code skills directory (see the Installation section above).
helpercraft is primarily written in HTML. It is open-source under Helpercraft on GitHub, so you can review or fork the full source.
Yes. SkillsLLM lists many other AI Agents skills you can browse and compare side by side. Open the AI Agents category from the badge at the top of this page, or use the Related Skills and comparison links further down to weigh helpercraft against similar tools.
No comments yet. Be the first to share your thoughts!
Unlocks once the catalog security scan passes (runs nightly).
⚠️ Third-Party Software Notice
This skill is third-party open-source software developed and hosted independently on GitHub. SkillsLLM is an informational directory and does not control or maintain the underlying repository.
Any security checks, ratings, or warnings displayed by SkillsLLM are automated and limited in scope. They do not constitute a security certification or guarantee that the software is safe, error-free, or free from malicious code, vulnerabilities, compromised dependencies, or prompt-injection risks.
Review the source code, permissions, dependencies, and configuration before installing or running any third-party skill. Use is at your own risk. To the maximum extent permitted by applicable law, SkillsLLM is not liable for losses arising from third-party software.
The deep catalog scan for this skill is still queued. Run an instant dependency check now instead.
Want it in every chat? On a computer, paste it into a new Gem at gemini.google.com. The Gem then shows up in your phone's Gemini app too.
Using the Claude app instead? Download an agent's zip and add it under Customize → Skills.
Agents are powerful, but as plain-text files they're easy to lose track of. I built Helpercraft to make creating and keeping them visual and friendly, so your team feels more like company than a pile of files. And you stay in control: your AI is told to ask before it puts your agents to work.
A short plain-text file: its job, how it talks, what it helps with and its rules. You can read it before you use it.
If you installed it as a skill, the AI uses it when your request matches its job. Ask for it by name or job, or save it as a Gem, a custom GPT or a Claude Project.
No. A job sets tone and safety rules, not expert knowledge. The agent is told not to diagnose, prescribe or give legal advice, and when to hand over to a person, but AI can ignore rules: in my test, asked for "just the number", it still gave a common label dose.
Only how the agent talks and works, never facts, numbers or safety rules. None of it is science. "MBTI" and "Myers-Briggs Type Indicator" are trademarks of The Myers-Briggs Company; Helpercraft isn't affiliated with or endorsed by them.
Yes: choose who can copy it in the Craft step, then share the file. Only use agents from people you trust, because your AI will follow their instructions.
Tested on 90 everyday tasks in Claude Code with Claude Sonnet, each pair of answers judged blind by an AI (a person's blind ratings in an earlier round agreed with it). Helpercraft 1.0.5 came out ahead of an AI that starts agents itself, slightly ahead of the same AI told to ask first (not a clear difference), and even with the same agents installed as skills. So it may improve answers a little in this setup; with other tools, models or settings, results may differ. Tests 3 to 6 · Tests 1 and 2
Your agents live in one folder instead of being copied into every project. Any AI tool that can read files can use it:
my-team/
├── START-HERE.md ← the AI reads this first
└── agents/
├── nurse-agent/SKILL.md
├── ui-designer/SKILL.md
└── cafe-assistant/SKILL.md
Copy for my AI copies the sentence to start with:
My agent team is in my-team. Read START-HERE.md there first, then suggest which of my agents should work on my task, and wait for my OK.
START-HERE.md asks the AI to:
If no agent comes close, it recommends helping you directly, then offers to save a new agent once your task is done. Asking first is an instruction, not a lock.
START-HERE.md and one agents/<name>/ folder per agent, with its SKILL.md and a small .helpercraft id file, so Helpercraft changes only its own folders.