Hysen Labsawesome-openclaw-skills
The awesome collection of OpenClaw skills. 5,400+ skills filtered and categorized from the official OpenClaw Skills Registry.🦞
Awesome OpenClaw Skills: thousands of agent skills, filtered
ClawHub holds thousands of community-built skills for OpenClaw, and this list keeps the ones that passed its filters: roughly 5,400 skills after removing spam, duplicates, weak descriptions, finance tools, and flagged malware.
What the list is
The repository is a collection of OpenClaw skills, filtered and categorized from the official OpenClaw Skills Registry, with over 5,400 skills. OpenClaw is a locally-running AI assistant that operates directly on your machine, and skills extend its capabilities so it can interact with external services, automate workflows, and handle specialized tasks. The collection is meant to help you discover and install the right skills, and it can also serve as inspiration for OpenClaw use cases. Skills come from ClawHub, the public skills registry, and are categorized for easier discovery.
What got filtered out
A table lays out the cleanup. Excluded were possibly spam, bulk accounts, bot accounts, test or junk entries at 4,065, duplicates or similar names at 1,040, low-quality or non-English descriptions at 851, crypto, blockchain, finance, or trade at 886, and malicious skills identified by security audits published by researchers, excluding VirusTotal, at 373. In total, 7,215 items were not taken from the official skills registry. The filtering is the core argument for why the list is smaller and more useful than the registry itself.
How skills get installed
Installation paths are documented. Global skills live in ~/.openclaw/skills/ and workspace skills in /skills/. The ClawHub CLI works for registry-managed skill folders outside a full OpenClaw workspace. There is also a low-friction route: paste a skill's GitHub repository link into the assistant's chat and ask it to use the skill, and the setup happens automatically in the background.
The registry rule
Only skills already published on ClawHub make it onto the list. Links to personal repos, gists, or any other external source are not accepted, so a skill must be published to the registry first. Contributors are asked to include the ClawHub link for their skill in the pull request description, and the clawskills.sh listings are managed separately. The rule keeps the list aligned with the registry's own content.
The security caveat
A dedicated notice separates curation from auditing. Skills in the list are curated, not audited, and their maintainers can update, modify, or replace them at any time. Before installing anything, the README advises reviewing potential security risks and validating the source yourself. OpenClaw has a VirusTotal partnership for scanning, with the report available on each skill's ClawHub page. Agent skills can hide prompt injections, tool poisoning, hidden malware payloads, or unsafe data handling patterns, so the guidance is to review the source code and use skills at your own discretion. If something looks wrong, the README asks you to open an issue.
Community notes