From 5c6911b44dd1e577cf5b71295f161d92cb532686 Mon Sep 17 00:00:00 2001
From: weishen <ishenwei@gmail.com>
Date: Sat, 25 Apr 2026 20:02:49 +0800
Subject: [PATCH] Auto-sync: 2026-04-25 20:02

---
 wiki/index.md                              |  1 +
 wiki/sources/testing-evidence-collector.md | 53 ++++++++++++++++++++++
 2 files changed, 54 insertions(+)
 create mode 100644 wiki/sources/testing-evidence-collector.md

diff --git a/wiki/index.md b/wiki/index.md
index a1dae050..2801302f 100644
--- a/wiki/index.md
+++ b/wiki/index.md
@@ -4,6 +4,7 @@
 - [Overview](overview.md) — living synthesis
 
 ## Sources
+- [2026-04-25] [Testing Evidence Collector Agent Personality](sources/testing-evidence-collector.md)
 - [2026-04-25] [Test Results Analyzer Agent Personality](sources/testing-test-results-analyzer.md)
 - [2026-04-25] [Performance Benchmarker Agent Personality](sources/testing-performance-benchmarker.md)
 - [2026-04-25] [Testing Reality Checker](sources/testing-reality-checker.md)
diff --git a/wiki/sources/testing-evidence-collector.md b/wiki/sources/testing-evidence-collector.md
new file mode 100644
index 00000000..0aea074c
--- /dev/null
+++ b/wiki/sources/testing-evidence-collector.md
@@ -0,0 +1,53 @@
+---
+title: "Testing Evidence Collector Agent Personality"
+type: source
+tags: []
+date: 2026-04-25
+---
+
+## Source File
+- [[Agent/agency-agents/testing/testing-evidence-collector.md]]
+
+## Summary（用中文描述）
+- 核心主题：EvidenceQA —— 一个以截图为核心证据的 QA Agent 个性化角色定义
+- 问题域：如何对 AI Agent 生成的前端实现进行严格的质量评估，避免"幻想式报告"（Fantasy Reporting）
+- 方法/机制：通过 Playwright 自动化截图 + 视觉对比 + 强制默认找问题（至少 3-5 个）来实现真实性检验
+- 结论/价值：QA 质量评估必须基于视觉证据，零问题报告是红色警报，必须强制提供截图
+
+## Key Claims（用中文描述）
+- EvidenceQA 相信"截图不会撒谎"——视觉证据是唯一可靠的真理
+- 首次实现总是存在至少 3-5 个问题，"零问题"是红色警报
+- 每个声明都需要截图证据支撑，无证据的声明视为"幻想"
+- luxury/premium 等描述词无截图支撑即为违规
+- 质量评级默认 FAILED，除非压倒性证据证明通过
+
+## Key Quotes
+> "Screenshots Don't Lie" — Visual evidence is the only truth that matters
+> "Default to Finding Issues" — First implementations ALWAYS have 3-5+ issues minimum
+> "Zero issues found" is a red flag - look harder
+> "Your job is to be the reality check that prevents broken websites from being approved"
+
+## Key Concepts
+- [[Visual Evidence]]：QA 评估的唯一可靠依据，通过 Playwright 自动化截图捕获
+- [[Fantasy Reporting]]：指无视觉证据支撑的声称，如"零问题"、"Luxury 级别"等
+- [[Reality Check Commands]]：强制性初始检查命令，包括 Playwright 截图、文件检查、grep 特征搜索
+- [[Specification Compliance]]：将实际截图与原始规范逐字对比，不添加规范外的额外要求
+- [[Accordion Testing Protocol]]：通过 before/after 截图对比验证手风琴组件的展开/折叠功能
+- [[Form Testing Protocol]]：验证表单提交、校验、错误信息展示的完整性
+- [[Mobile Responsive Testing]]：在 desktop/tablet/mobile 三种分辨率下验证布局和导航
+
+## Key Entities
+- [[EvidenceQA]]：截图驱动型 QA Agent，以视觉证据为唯一真理，默认发现 3-5+ 问题
+- [[Playwright]]：自动化截图工具（qa-playwright-capture.sh），提供 comprehensive screenshots 和 test-results.json
+
+## Connections
+- [[Testing Reality Checker]] ← related_to ← [[Testing Evidence Collector]]
+- [[Testing Test Results Analyzer]] ← related_to ← [[Testing Evidence Collector]]
+- [[Testing Performance Benchmarker]] ← related_to ← [[Testing Evidence Collector]]
+
+## Contradictions
+- 与声称"零问题"的报告冲突：
+  - 冲突点：首次实现的问题数量
+  - 当前观点：默认发现 3-5+ 问题，"零问题"是红色警报
+  - 对方观点：声称"零问题"即通过
+```