WeChat Mini Program Cloud Testing

Problem

WeChat Mini Program cloud testing often involves repeating browser actions: submitting tasks, checking status, downloading reports, managing cases, and diffing baseline images. These steps are hard to embed in CI when they require switching between Android/iOS devices, online/experience/dev builds, and Monkey, Minium, replay, or AI cases. This skill exposes the cloud testing workflow as HTTPS calls to https://minitest.weixin.qq.com/thirdapi/, so test execution and result retrieval can be scripted and orchestrated.

How it works

Before calling any endpoint, provide token and group_en_id. The token authenticates the user, while group_en_id identifies the target project or product in the cloud testing platform. A typical flow starts with POST /plan to submit a test task, then polls GET /plan or GET /plan_list for status. After completion, use GET /share_url and GET /report/static_resource to obtain a shareable link and report assets. The skill also covers test plans, cases, image comparison, and SourceMaps: create Minium or AI test plans, upload/download ZIP cases, list Minium/replay/AI cases, set baseline images and re-diff them, and upload SourceMap packages. Quality inspection tasks are submitted and queried through the /union_plan endpoints.

Boundaries and cautions

This skill is an external-service integration, so all requests go to minitest.weixin.qq.com and depend on the caller's real permissions, quotas, and project configuration. token is a credential; avoid hard-coding it in repositories or logs. group_en_id must match the intended project, otherwise tasks and cases may be created in the wrong project. Some features such as AI Monkey, fast testing, and special cloud capabilities may depend on paid access or platform switches, so use parameters like device_ids, ai_monkey, and special_cloud only when the project is entitled to them.