engineering-practices

Comprehensive collection of essential software engineering practices and recommendations.

A centralised hub for learner around the globe from A-Z. You can find collections of manuals, blogs, hacks, one liners, courses, other free learning-resources and more

A responsive and robust exercise repository for daily workout

Software Engineering Unlocked Podcast

Development Guidelines

Specwright is a codebase-aware engineering platform for teams that want AI to operate inside architecture, policy, and evidence boundaries instead of around them.

One copy of every cross-repository rule for AI-agent sessions, mounted as a git submodule in five repos and enforced by CI pin checks.

Architecture decision records, system design case studies, and delivery templates demonstrating pragmatic SaaS architecture, cloud trade-offs, and consulting-grade technical leadership.

Practical workflow notes for AI-assisted coding agents

Clean-first engineering discipline as an agent skill, in English and Chinese: root-cause repair, one source of truth, no residue

Adapting timeless software engineering ideas into skills, prompts, workflows, and agentic development practices.

Source-backed engineering practices and a reviewed shelf of skill repos, as a Claude Code mod and a CLI.

A practical reference template for building AI Agent-friendly software repositories.

A rulebook for letting an AI coding agent write code in a repository you are responsible for, with commands that measure a repository against the rules.

For now, just a personal list of 📚 and publications

An open handbook of everything a software engineer needs to know.

Documentation and Guidelines for using atc-net components

Field notes on long-horizon AI agent collaboration: 20+ measured failure modes, 12 engineering guidelines, and a portable 5-layer mechanism design. Installer + commit hook, no vendor lock-in.

Evidence-first engineering discipline for Claude Code. Guards that ship with their calibration, a knowledge base that lints itself, and a plan gate that refuses unproven premises.

A small, opinionated set of reusable AI agent skills for disciplined, low-ceremony software development. Designed to help humans and AI work together safely in greenfield and legacy codebases—without tickets, workflows, or process bloat.