• Currently building bilingual QA resources, evidence-driven AI quality tools, local-first Quality Engineering workbenches, and learning products around AI-Native-QA-Workbench, ai-automation-test-starter, awesome-qa-skills, ai-native-qa-agents, QCov, ai-test-auditor, dsh-qa, and naodeng.com.cn
• Also building practical web products such as tidemail and starry-island
• Focused on AI-native quality engineering, quality evidence, test quality auditing, and practical testing knowledge sharing
• I regularly write articles on https://inaodeng.com
• How to reach me: [email protected]
• Twitter: @inaodeng

• Building AI-Native-QA-Workbench as a local-first Quality Engineering workbench that connects requirements, quality risks, tests, execution, evidence, and human-controlled decisions; v0.3 now adds the event-driven Quality Engineering Loop, deterministic QualityAssessment/QualityGate evaluation, and an explicit HumanDecision boundary
• Developing ai-automation-test-starter with a v0.1 deterministic Pytest API vertical slice and framework-native Learn, Build, Verify, and Ship flows
• Advancing ai-native-qa-agents through v0.4 Test Effectiveness & Mutation, with bounded mutation evidence, effectiveness scoring, trace persistence, and quality gates
• Shipping quality evidence tools: QCov reached v1.0 with change-to-obligation impact analysis and adapter SDKs, while ai-test-auditor v1.2.1 keeps source-only JavaScript/TypeScript auditing and continues to improve static audit precision and coverage
• Evolving awesome-qa-skills through v1.6 Phase 1 Match Review and Evidence-Bounded Governance, with 330 bilingual skills, 10 workflows, 150 testing types, and evidence-bounded evaluation contracts
• Building dsh-qa as a QA workbench plugin for DeepSeek Harness; v0.7.2 supports the dsh-v0.2.0-rc.2 host, restores project-local DSH history, and documents the bilingual user guides
• Building reusable Codex Skills for structured workflows, including Markdown-first learning plans, release delivery, version DoD evidence, and the source-backed bilingual ai-qa-weekly digest
• Maintaining bilingual QA skills and prompts for Codex, Cursor, Claude Code, Kiro, OpenCode, and Trae
• Growing inaodeng.com as a bilingual testing site with structured learning paths, glossary content, bilingual QA workflow series, and weekly testing articles
• Shipping tidemail v0.2.0 as a bilingual Cloudflare Workers/D1 temporary email service with responsive UI and resilient inbox refresh behavior
• Building starry-island v0.2.0, a learning platform for ages 7–12 with 158 grade 1–3 Chinese, math, and English curriculum items and reviewed practice content

• Where AI Actually Helps in Testing—and Where It Only Looks Like It Does
• Circuit Breaker Testing Skill: Verify open, half-open, and recovery behavior under dependency failure
• Dependency Failure Testing Skill: Make timeout, stale data, and partial failure observable
• Agent Long-Running Testing Skill: Keep progress, leases, and cancellation visible during long jobs
• Multi-Agent Testing Skill: Check handoffs, authority, and disagreement across agent roles
• Agent Loop Testing: Measure Progress and Stop Safely

View all articles

Testing Tools

Languages

AI & Prompting

Web & Docs

DevOps

An AI-native, local-first Quality Engineering workbench that connects requirements, quality risks, tests, executions, and evidence into traceable quality decisions. The v0.3 release adds an event-driven Quality Engineering Loop, deterministic QualityAssessment/QualityGate evaluation, explicit HumanDecision records, a local API, and a bilingual UI.

A beginner-first, AI-native test automation kit. Its v0.1 Pytest API slice keeps framework-native behavior and connects Learn, Build, Verify, and Ship with deterministic checks and explicit evidence boundaries.

An evidence-driven reference architecture for AI-native quality agents, with controlled runtimes, deterministic analysis, bounded model reasoning, and verification gates across the delivery lifecycle. v0.4 adds a bounded Test Effectiveness & Mutation workflow with traceable mutation evidence.

A Quality Evidence Gap Engine that shows what tests and reports still do not prove about a requirement, risk, or code change. v1.0 adds change-to-obligation impact analysis and a public adapter SDK.

A local-first CLI that audits JavaScript and TypeScript test source for deterministic FAKE and WEAK signals before false confidence ships. The current stable release is v1.2.1.

A bilingual AI-native Quality Engineering skill system with 330 Skills (zh+en), 10 workflows, 150 testing types, and 5 skill-engineering skills. The v1.6 release adds Phase 1 Match Review and Evidence-Bounded Governance.

A local QA workbench plugin for DeepSeek Harness, covering AI-assisted requirements, cases, defects, milestones, reports, kanban, and calendar views. v0.7.2 supports the dsh-v0.2.0-rc.2 host and restores project-local DSH conversation history.

A bilingual testing blog and learning site built around automation practice, testing encyclopedia content, bilingual QA workflow series, and searchable QA resources. It is built with Astro 7.x and deployed on Cloudflare Workers.

A bilingual QA Prompt Baseline covering 256 capabilities and modules across requirements, test design, automation, performance, security, release, production quality, and AI testing.

A lightweight bilingual temporary email service built with Cloudflare Workers and D1. v0.2 improves responsive layouts, localized legal pages, and cached inbox behavior when a refresh fails.

A playful learning platform for ages 7–12 with interactive practice across Chinese, math, and English. v0.2 aligns and reviews grade 1–3 curriculum content and adds course-quality checks.

If you find my work helpful:

Thank you to the following supporters for their kindness:

_18773050607

_{Made with love by Nao Deng}