← Back to Skills Marketplace
luowanqian

Skill Creator

by Luo Wanqian · GitHub ↗ · v0.1.0 · MIT-0
cross-platform ✓ Security Clean
316
Downloads
0
Stars
0
Active Installs
1
Versions
Install in OpenClaw
/install skill-creator-3
Description
Create new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, edit, or optimize a...
Usage Guidance
This skill appears to do what it says: create, evaluate, and iterate on skills using the included scripts. Before installing/using it, consider: 1) The scripts recursively read and embed files from the workspace (text, images, PDFs, xlsx, binaries). Do not run this on a workspace containing secrets you don't want embedded into an HTML file or exposed. 2) generate_review.py may call lsof and attempt to kill processes on a port and will open a browser — ensure you are comfortable with that behavior on your machine. 3) Review the included Python files (especially eval-viewer/generate_review.py and any script that reads files or calls subprocess) if you are unsure; running the code in an isolated test workspace is recommended. Overall the package is internally coherent for its stated purpose, but treat it like any local tooling that reads and serves files: inspect inputs and run in a safe environment if you have sensitive data.
Capability Analysis
Type: OpenClaw Skill Name: skill-creator-3 Version: 0.1.0 The skill bundle is a comprehensive developer toolset for creating, evaluating, and optimizing AI agent skills. It includes Python scripts for benchmarking, description optimization via the `claude` CLI, and a local web-based viewer for reviewing results. While the bundle utilizes high-risk capabilities such as subprocess execution and local server hosting, these behaviors are transparently documented and strictly serve the tool's purpose of skill development. No evidence of malicious intent, data exfiltration, or unauthorized system modification was found across the code or agent instructions.
Capability Assessment
Purpose & Capability
Name/description match the delivered artifacts: SKILL.md describes creating and iterating on skills and the bundle contains agents (analyzer, comparator, grader), evaluation scripts, an eval viewer HTML, and packaging/benchmarking scripts. There are no unrelated environment variables, binaries, or external credentials requested, so the requested capabilities align with the stated purpose.
Instruction Scope
The runtime instructions direct the agent to run the included Python scripts (e.g., eval-viewer/generate_review.py) and to read and present evaluation results. Those scripts recursively read workspace directories, embed file contents (text, images, PDFs, xlsx, other binaries) and serve a self-contained HTML page. This is expected for an eval/skill-creation tool, but it means the skill will access arbitrary files in the workspace (except some named skips).
Install Mechanism
There is no install spec and the skill is instruction-plus-scripts only (no external downloads). All included code is present in the bundle and uses only standard Python stdlib APIs; no third-party package installs are required. This is proportionate and low risk in terms of installation.
Credentials
The skill declares no required env vars or credentials (good). However, the scripts embed and base64-encode files from the workspace and may call subprocesses (e.g., lsof, os.kill) and open a browser. While these actions are consistent with a local review server, they give the skill broad read access to any files placed in the workspace and can interact with local processes/ports — review what data will be present in the workspace before running.
Persistence & Privilege
The skill is not always-enabled and does not request system-level persistence. It saves feedback.json and other artifacts to the workspace (expected for an eval tool) but does not modify other skills or global agent configuration.
How to Use
  1. Make sure OpenClaw is installed (local or Docker)
  2. Run the install command in chat: /install skill-creator-3
  3. After installation, invoke the skill by name or use /skill-creator-3
  4. Provide required inputs per the skill's parameter spec and get structured output
Version History
v0.1.0
Initial release of Skill Creator: an end-to-end tool for creating, editing, and optimizing skills, with evaluation and benchmarking support. - Guides users through the full lifecycle: intent capture, drafting, evaluation, and iterative improvement of skills - Provides detailed best practices for writing and organizing skill files, including SKILL.md formatting and resource bundling - Includes advice for communicating effectively with users of varying technical backgrounds - Supports benchmarking and quantitative/qualitative evaluation workflows for skill quality assessment - Offers processes for optimizing skill descriptions to improve triggering accuracy
Metadata
Slug skill-creator-3
Version 0.1.0
License MIT-0
All-time Installs 0
Active Installs 0
Total Versions 1
Frequently Asked Questions

What is Skill Creator?

Create new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, edit, or optimize a... It is an AI Agent Skill for Claude Code / OpenClaw, with 316 downloads so far.

How do I install Skill Creator?

Run "/install skill-creator-3" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is Skill Creator free?

Yes, Skill Creator is completely free, licensed under MIT-0. You can download, install and use it at no cost.

Which platforms does Skill Creator support?

Skill Creator is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created Skill Creator?

It is built and maintained by Luo Wanqian (@luowanqian); the current version is v0.1.0.

💬 Comments