AGENT SKILLS

Agent Skills Library

A curated directory of third-party Agent Skills, organized by use case, platform and safety rating. Before installation, review package permissions, dependencies and runtime risks.

Found 5 Agent Skills

cls-certify

The enterprise-level Agent Skill security certification tool uses six-dimensional in-depth detection technology (static analysis, dynamic behavior, dependency audit, network traffic, privacy compliance, threat intelligence). It is a personal developer community project and can output S+/S/A/B/C/D security level assessment and visual reports.

Security & Sandbox S rating

OpenClaw Claude Code Cursor Codex

docker-sandbox

The document-based skills of Docker's official sandbox function guide users to create a VM-level isolation environment to safely run AI Agents, achieving zero-trust code execution and fine-grained network control.

Security & Sandbox A rating

OpenClaw Claude Code Cursor Codex

flatnotes-tasksmd-github-audit

Based on GitHub as the source of truth, it automatically audits the consistency of Flatnotes knowledge base and Tasks.md dashboard, detects outdated notes, missing links and project drift, and generates executable repair plans.

Security & Sandbox B rating

OpenClaw Claude Code Cursor Codex

clawdbot-backup

ClawdBot's official ecological backup tool implements cross-device configuration synchronization and version control through tar/git/rsync, ensuring the safe migration of user skills and assets.

Security & Sandbox S rating

OpenClaw Claude Code Cursor Codex

db-readonly

A secure read-only query tool based on the whitelist mechanism that manages credentials through environment variables to achieve zero-risk data exploration and report generation.

Security & Sandbox A rating

OpenClaw Claude Code Cursor Codex