Skill Creator (Claw)

Trust Score 85/100

Create, test, and iteratively improve OpenClaw skills; includes eval workflows, test-case guidance, and packaging tools.

triggers:create skillskill creatorauthor skilloptimize skill descriptionskill eval

GitHub SKILL.md

What it does

Guides authors through drafting SKILL.md files, creating eval test cases, running with/without-skill comparisons, and packaging final skills. It provides a structured loop for iterative improvement and description optimization.

When to use it

Use this skill when you want to author a new skill, optimize an existing SKILL.md, run trigger evaluation, or benchmark skill behaviour against an external Agent. Ideal for maintainers and power users looking to improve skill triggering and reliability.

What's included

Scripts: includes scripts/ and an eval-viewer for packaging and reviewing test runs.
References: guidance docs and example templates are bundled in references/.
Instructions: detailed authoring workflow with test-case and benchmarking guidance.

Compatible agents

Targets environments with subagent or command-line capability (Claude Code, Cowork-style setups) where running parallel evals and generating review HTML is feasible.

Audit Summary

Skill Creator (Claw) is a comprehensive skill authoring and evaluation framework that guides creating, testing, iterating, and optimizing OpenClaw skills. It includes an eval viewer, benchmark aggregation, description optimization loop via external Agent API, and packaging tools. Scripts are well-written but 4 of 7 non-trivial scripts fail to run standalone due to using 'from scripts.xxx' import paths instead of relative imports, requiring them to be run as modules (python -m scripts.xxx). The two utility modules (api_client.py, utils.py) import cleanly.

Watch Out

Scripts use 'from scripts.xxx import' — must be run as 'python -m scripts.script_name' from the skill root, not as standalone scripts
Requires external Agent API configuration (claw-config.json or CLAW_API_URL/CLAW_TOKEN env vars) for eval and description optimization features
api_client.py and improve_description.py contain Chinese comments alongside English

Notes

Well-structured and genuinely useful skill. The import path issue is the main code quality ding — scripts should use relative imports or be runnable standalone. Security is solid: no hardcoded secrets, no destructive commands, no suspicious network calls. API calls go only to user-configured endpoints.

Information

Repository: skill-creator-claw
Stars: 9

Trust Score

Overall85

Security92

Code Quality78

Architecture80

Usefulness85

Related Skills

Development Worktree

Create an isolated git worktree for feature work, auto-run project setup, and verify a clean test baseline before development.

Readwise Reader Document Management

Manage Readwise Reader documents: list, save, search, move, tag, highlight, export and bulk-edit via official and custom CLIs.

Full Stack Builder

End-to-end builder that scaffolds, implements, tests, and optionally deploys web and API applications from a natural-language specification.

Godot WeChat Minigame Adapter

Apply a version-locked adapter and patch bundle to port an official Godot checkout to WeChat Mini Game, handling WXMEMFS, wasm packaging, audio fixes, and runti

ezBookkeeping API Tools

Command-line API tools for ezBookkeeping: record and query transactions, retrieve accounts/categories/tags, and fetch exchange rates for self-hosted personal fi

Claw Bench

Benchmarking skill that guides an agent through a structured suite of capability tests and reporting steps for leaderboard submission.

Tmux Bridge

Give an agent controlled access to a local tmux terminal bridge so it can run shell commands, capture output, and manage sessions on the user's machine.

SourceSage CLI

Generate concise, LLM-friendly repository summaries (Markdown) using the SourceSage CLI — supports lite mode, language switching, and targeted repo analysis.

Back to Skills