AI-assisted software development
Generate and review code with AI, check architecture and implementation, plan refactors, and evaluate tests and regression risks.
Start here
Begin with these resources. The sections below organize the rest of the topic by subject.
-
Article
Vibe Coding Risks: Code Defects and Verification Gaps
Understand the verification gaps that can remain even when AI-generated code appears to work.Read article -
Guide
How to Choose the Right Code Review Method
Choose a review method for architecture, implementation, code quality, or tests.Open guide -
Skill
Reviewing Code Against Language and Framework Best Practices
Open a reusable Skill for reviewing supplied code against relevant engineering guidance.Open skill
Review architecture and implementation
Review module boundaries and dependencies, check code against official guidance, or run both as one quality gate.
-
Guide
Software Architecture Boundary Review: Step-by-Step Guide
Review software layers, module boundaries, ownership, dependency direction, interfaces, and coupling using evidence from inspected files and diagrams.Open guide -
Guide
Review an Implementation Against Official Docs and Standards
Step-by-step workflow for checking code, configuration, API usage, or design materials against official documentation, standards, and vendor guidance.Open guide -
Guide
Engineering Quality Gate for Architecture and Code Review
A focused procedure for reviewing architecture boundaries and implementation quality against official technical guidance.Open guide
Tests and regression risk
Check what a change could break, find gaps in existing tests, and write new unit tests.
-
Guide
Test Coverage and Regression Review After Code Changes
Review whether code, configuration, API, fixture, mock, or test changes are covered by tests and protected against regression.Open guide -
Skill
Reviewing Regression Risk
Identifies what a specific code, API, schema, configuration, dependency, integration, or runtime-behavior change could break.Open skill -
Skill
Reviewing Test Coverage
Audits existing tests against the features, behavior, requirements, or implementation they should cover.Open skill -
Skill
Generating Unit Tests
Writes focused unit tests that follow the project's existing test conventions, runs them when possible, and reports the actual result.Open skill
Write and restructure code
Generate code a beginner can follow, or plan a refactor that keeps behavior unchanged.
-
Prompt Template
Write Beginner-Friendly Code
Check the goal, missing context, and assumptions before writing minimal code a beginner can follow.Open prompt template -
Skill
Planning Behavior-Preserving Refactors
Plans a small, staged refactor that improves code structure, readability, or maintainability while preserving externally visible behavior.Open skill
Where coding workflows belong
Decide when a coding workflow belongs in ChatGPT, Claude, or an API integration.
-
Reference
ChatGPT Setup: Choosing Projects, Skills, Plugins, and Tools
Choose models and place workflows across Chat, Work, Codex, Projects, Skills, Plugins, existing GPTs, Scheduled Tasks, memory, and Library.Open reference -
Reference
Claude Projects, Skills, or Claude Code: What to Use When
Use this reference to decide where a workflow belongs in Claude: Projects, Skills, Claude Code, MCP, subagents, artifacts, or runtime prompts.Open reference -
Reference
AI API and Agent Architecture: Where Instructions, State, and Tools Belong
Use this reference to decide when an AI workflow belongs in an API, internal agent, backend service, retrieval system, tool workflow, or production automation.Open reference