Openclaw Ticket Assistant

OpenClaw工单助手，根据客服截图自动分析客户问题并创建工单。触发词：工单、创建工单、客服截图、提工单。

This skill is coherent with its goal of creating internal tickets from support screenshots, but check these before installing: - Transparency: SKILL.md requires the agent to hide intermediate steps; consider whether you want an automated agent to run silently (that could hide mistakes or unintended actions). - Login handling: the skill will instruct you to scan a QR code to log into your corporate system and will capture the logged-in account from the page. Confirm whether that account info is stored persistently by the platform and whether that storage is acceptable under your policies. - Data access: the skill reads resources/customers.json (contains many customer IDs). Verify the customer list is appropriate and doesn't expose sensitive data you wouldn't want available to the agent. - Internal endpoints: the skill automates an internal URL (https://t.aliyun-inc.com/workbench/list). Ensure the agent's automated browser access is allowed by your organization and that no additional credentials are required beyond the user's scan login. - Missing files note: README references a scripts/ directory that isn't present in the package—ask the author for full source or clarification if you expect extra helper scripts. If you decide to proceed, test in a controlled environment (non-production) first and confirm how/where the agent stores any recorded login/account information.

Type: OpenClaw Skill Name: openclaw-ticket-assistant Version: 1.0.0 The skill is a productivity tool designed to automate ticket creation on an internal Alibaba ticket system (aliyun-inc.com) by analyzing customer service screenshots. It features a well-defined workflow including customer identification via a local JSON file (customers.json), automated form filling, and a mandatory user confirmation step (Step 3.3) where a screenshot of the prepared ticket is shown before submission. Authentication is handled safely by guiding the user through a QR code login process (login-guide.md) rather than requesting credentials, and no evidence of data exfiltration or unauthorized remote execution was found.