v0.5 — 四层 QA Skills 体系与双语治理 / Four-Layer QA Skills and Bilingual Governance

中文

概述

v0.5 将项目演进为以研发与测试阶段组织的四层质量能力体系：

Core QA Skills → Engineering QA Skills → Production Quality Skills → AI Native QA Skills

仓库现提供 78 项质量能力（中英文共 156 个可独立安装的 Skill 包），并以中文为主、英文镜像的项目级文档方式维护。

主要更新

• 新增并完善 Engineering QA、Production Quality 与 AI Native QA 能力，包括变更影响、契约、测试数据、回归选择、性能分析、生产验证、事故/Trace/指标分析，以及 AI、LLM、Prompt、Agent 与注入防护测试。
• 以研发与测试阶段重组 README 导航，提供四层能力、工作流和专项 Skill 的清晰入口。
• 新增 3 个 Skill Engineering 能力：变更验证、Skill 文案契约审查与过程性文案清理。
• 建立项目级中英文文档治理：中文主文档、英文镜像、双向切换和自动校验。
• 新增 Skills 关系图、全量索引、演进路线图，以及 GitHub Actions 项目质量门禁。
• 按 skill-review 重新审计全部 65 个测试类型主题（中英文 130 个包）；发现已记录在双语审计报告中，未掩盖待维护事项。

质量验证

• 156 个 Skill 的元数据、独立安装、完整性和外部快照检查通过。
• 156 份 skill-up eval YAML 校验通过。
• 中英文项目文档、链接、目录配对和 README 分类检查通过。
• GitHub Actions 在 PR 与 main / explore 推送时运行项目质量门禁。

已知维护项

审计发现 1 项 Medium 与 1 项 Low：性能测试速查中的一条 k6 场景命令需要替换，Selenium 示例依赖需要在隔离环境复核后更新。详见 全部测试类型 Skill 审计报告。

---

English

Overview

v0.5 evolves the project into a four-layer quality-capability model organized around delivery and testing stages:

Core QA Skills → Engineering QA Skills → Production Quality Skills → AI Native QA Skills

The repository now provides 78 quality capabilities (156 independently installable Chinese and English Skill packages), with Chinese-first project documentation and complete English mirrors.

Highlights

• Added and strengthened Engineering QA, Production Quality, and AI Native QA capabilities: change impact, contracts, test data, regression selection, performance analysis, production verification, incident/trace/metric analysis, and AI, LLM, Prompt, Agent, and prompt-injection testing.
• Reorganized README navigation around the four capability layers, delivery stages, workflows, and specialist Skills.
• Added three Skill Engineering capabilities: change verification, prose-contract review, and process-prose trimming.
• Established project-level bilingual documentation governance: Chinese primary documents, English mirrors, bidirectional language switches, and automated validation.
• Added a Skills relationship graph, complete catalog, evolution roadmap, and a GitHub Actions project quality gate.
• Re-audited all 65 testing-type topics (130 Chinese and English packages) with skill-review; findings are recorded transparently in bilingual audit reports.

Quality Validation

• Metadata, standalone-installation, integrity, and external-snapshot checks passed for all 156 Skills.
• All 156 skill-up eval YAML configurations passed validation.
• Project-document bilinguality, links, directory pairing, and README classification checks passed.
• GitHub Actions runs the project quality gate on pull requests and pushes to main and explore.

Known Maintenance Items

The audit records one Medium and one Low item: one k6 scenario quick-reference command needs replacement, and Selenium example dependencies should be refreshed after isolated-environment verification. See the All Testing-Type Skills Audit Report.