创建或卸载独立的 OpenClaw Agent，支持独立工作区、可选通道绑定（飞书/微信等）、模型选择。触发词：创建agent、new agent、创建机器人、agent setup、new bot、卸载agent、删除agent、remove agent

This skill roughly does what it says (create/uninstall OpenClaw agents) but has red flags you should consider before installing or running it: 1) It references and edits local configuration and workspaces and asks you to provide channel credentials — only proceed if you trust the skill and its operator. 2) It uses a hardcoded filesystem path (/Users/honor/.qclaw/openclaw.json) and an application path under /Applications; verify these paths are correct for your system and not targeting another user's files. 3) The skill instructs direct edits (Python) to remove keys because gateway config.patch 'cannot delete keys' — prefer tools that safely remove keys or inspect the wrapper script and gateway behavior before running destructive operations. 4) Ask for the qclaw-openclaw wrapper script contents and verify it calls the real CLI safely; test operations in a backup or disposable environment first. 5) Do not supply channel credentials unless you know how they are stored and are prepared to rotate them if needed. If you want to proceed, request the owner/source code of the wrapper and confirm the paths and backup/rollback behavior.

Type: OpenClaw Skill Name: agentcreate Version: 1.0.4 The skill bundle contains hardcoded absolute paths to a specific user's home directory (/Users/honor/.qclaw/openclaw.json in references/delete.md), which is a significant security and portability flaw. It also includes instructions for the AI agent to execute Python code to directly modify configuration files and handles sensitive credentials like Feishu appId and appSecret. While these actions are aligned with the stated purpose of managing OpenClaw agents, the lack of path sanitization and the direct file manipulation represent high-risk behaviors.