multica-test-t3-ui-automation · diff

v1.2.0 to v1.2.0

29 added, 85 removed. Audit A to A.

---
name: multica-test-t3-ui-automation
- description: T3 UI 自动化：pytest+Playwright，1 功能用例 = 1 个 e2e 测试；路径读 MULTICA.md §2；在 deploy URL 执行。
+ description: T3 UI automation—only after G2.5 PASS; run pytest+Playwright 1:1 by MULTICA §2 e2e path and T1 cases; join overall verdict with API side. No local mock.
version: 1.2.0
+ metadata:
+ orchestrates:
+ - multica-test-orchestration
---
- # Test UI Automation（功能用例 → UI 1:1）
+ # UI Automation (T3 · UI)
- 把 **T1 Confluence / T2 入库** 的 Playwright 用例在 **部署环境** 执行。
+ T3 only after **G2.5 PASS**; an AC can only be verified by UI → run this skill alone; if it also has an API channel → run both, orchestration merges one report.
- > T2 应在 **MULTICA.md §2 UI E2E 路径** 编写/积累脚本（见 `docs/test-automation-in-repo.md`）。T3 门禁见 `multica-test-orchestration`。
+ > Integrated from `ui-automation-squad`. **Material assembly + BLOCKED see `multica-test-orchestration`.**
- ## Platform 协作
+ ## Required Items (stop + interruption list if missing; summary from comments)
- | Platform skill | 本 skill 用途 |
+ | Required | How it counts as complete |
| --- | --- |
- | — | 无直接 platform 调用；`page_url` / 用例集来自 Issue 评论与 T1 产物 |
-
- UI 在部署环境执行；环境 URL 由 DevOps / Leader 写入 JIRA 评论（见 `multica-artifact-cicd-sync`）。
-
- ## 原则
-
- - **入仓路径**：Read 目标 repo **MULTICA.md §2**（分仓按 Issue 矩阵各读一份）；缺文件 → BLOCKED
- - **无 e2e 目录时**：在 MULTICA 指定路径按 `framework-template.md` 初始化（禁止 `/tmp/ui-auto-*`）
- - **禁止前置**：`xmind-ui-automation`、`req-figma-ui-automation`、公司 **`playwright.git` 外挂仓库**（与本仓 MULTICA §2 路径不矛盾）
- - **1:1**：每条自动化 CASE → 恰好一个 `test_*.py`
- - **步骤对齐**：`steps[i]` → 第 i 个操作；`expected_results[i]` → 该步断言（等长）
- - **不编造**：locator 来自 **打开 page_url 后**的真实页面；找不到 → **BLOCKED**
-
- ## 输入（缺则中断）
-
- ```text
- repo_root: # Issue 仓库矩阵；Read MULTICA.md
- e2e_dir: # MULTICA.md §2 UI E2E 路径（默认 tests/e2e/）
- cases_source: # T1 Confluence 或 e2e_dir 内 test_*.py
- page_url / deploy_url: # Issue 评论
- jira_key:
- auth_mode: # none | form_login | sso | cookie
- auth_user_env / auth_pass_env:
- ```
-
- ### 账号密码（强制）
-
- - `page_url` **不等于**已登录；需登录系统必须给 `auth_mode` + env 变量名
- - **禁止**把账号/密码写进：用例 JSON、脚本、XMind、JIRA 评论、Allure 正文
- - 用例前置只写权限语义；真实凭据只放 Agent Secret / gitignore `.env`
- - 执行前检查 `auth_user_env`/`auth_pass_env` 非空；为空 → 中断，提示配置 Secret **勿回贴 Issue**
- - `auth_mode=none`：仅无需登录页面
- - `auth_mode=cookie`：本地 `storage_state` 路径，不把 cookie 贴评论
-
- 未确认、无 URL、需登录但 env 空、steps/expected 不等长 → 停止。
-
- ## 流程
-
- ### Phase A — 确认用例集
-
- 1. 只读 `functional`（忽略 `api`）
- 2. 校验 `len(steps)==len(expected_results)` 且非空
- 3. 列出 case_id → test 方法对照表（条数 = functional 条数）
+ | G2.5 PASS | Leader/DevOps explicit G2.5 PASS (or deploy_url + Leader written authorization) |
+ | deploy_url | Reachable environment URL |
+ | page_url | Page URL; if omitted must equal deploy_url (state in comments) |
+ | e2e script | **Path in MULTICA.md §2**; in deploy branch at G2.5 SHA, **1:1** with cases |
+ | ui_auth | `auth_mode=none` or "account via Secret variable name"; **never plaintext password from comments** |
- ### Phase B — 框架（e2e_dir）
+ **Never** use `xmind-ui-automation` / `playwright.git` / `req-figma-ui-automation` as prerequisites.
- 若 `{repo_root}/{e2e_dir}/` 无 `pytest.ini`：按 `framework-template.md` 生成于 **该目录**。
- 已有本 skill 骨架 → 增量加用例。
- 勿在 repo 外另建平行 playwright 工程。
+ ## Flow
```bash
- cd {e2e_dir} # MULTICA.md §2 UI E2E 路径
- python -m venv .venv
- .venv/Scripts/pip install -r requirements.txt
- .venv/Scripts/playwright install chromium
+ pytest <MULTICA.md §2 e2e path> --base-url <page_url> --alluredir ./reports/ui
```
- ### Phase C — 一对一生成/补全
-
- T2 应已建骨架；T3 补 locator、修断言。规则见 [`references/case-mapping.md`](references/case-mapping.md)。
+ Cases source: T1 Confluence t1-cases + T2 supplement. Missing script path → interruption list.
- ### Phase D — 执行
+ ## Output and merge
- ```bash
- cd {e2e_dir}
- .venv/Scripts/pytest . -q --alluredir artifacts/allure-results
+ ```markdown
+ # T3 UI Automation — {JIRA_KEY}
+ - G2.5: PASS @ {sha} deploy: {url} page: {page_url}
+ - e2e: {MULTICA.md §2 path}
+ - total / pass / fail / error
+ - per case: id / title / status / screenshot / log path
+ - conclusion: PASS | FAIL | BLOCKED
```
- ## 输出
-
- | 产物 | 说明 |
- | --- | --- |
- | `{e2e_dir}/` 变更 | **commit 到 deploy branch**（FIX 后同样） |
- | `artifacts/case_map.md` | case_id ↔ 文件 ↔ 结果 |
- | Allure / pytest 日志 | 按条 PASS/FAIL/SKIP(BLOCKED) |
- | **Issue 评论** | 映射表 + commit SHA + 失败明细 |
-
- 本 skill **不对整单 AC 做 PASS 裁决**；T3 总报告由 `multica-test-orchestration` 合并。
-
- ## FAIL / ERROR 分诊
-
- UI 自动化 FAIL 时：**@Tester** 按 `multica-test-orchestration/references/t3-failure-triage.md` 分诊。locator/步骤错 → Tester 修；页面行为不符 AC → 派 FrontendDev。禁止 Frontend 未分诊就改 pytest 降标。
+ `multica-test-orchestration` merges with API side: **one side only** verifies a given AC, missing that side → **whole ticket cannot be PASS**; errors → triage (`references/t3-failure-triage.md`).
- ## 禁止
+ ## Why it works
- - 使用 `xmind-ui-automation`、`req-figma-ui-automation`、**外部 playwright.git 替代本仓 MULTICA §2 e2e 目录**
- - 用例只存在于临时目录、不入 Git
- - 一条功能用例拆成多个 test 或多条合成一个
- - 把接口请求写进 UI 用例（接口走 Apifox）
- - G2.5 前把结果当成 T3 PASS
- - 把明文密码写进 Issue 或脚本
+ UI verification bound to the actual deployed page; 1:1 with cases avoids "runs but doesn't cover"; running after G2.5 avoids a green build masking a broken environment.