Workflows
AI Coding Workflows
Move from a vague request to a bounded task, isolated implementation, visible checks, reviewed merge, and recoverable release.
The workflow around an agent determines whether fast generation becomes maintainable software or merely faster rework.
Delegate a repository task
Give the agent enough context to act while keeping scope and permissions narrow.
A practical agent-first workflow covering task briefs, context, worktrees, permissions, tests, handoff, and human review.
Practical guideClaude Code + Cursor Setup for Reviewed Agent WorkA repeatable setup using Cursor for navigation and review, Claude Code for scoped tasks, and git worktrees for isolation.
Review generated code
Inspect behavior, access, failure modes, dependencies, tests, and the size of the change.
Review AI-written code by behavior and risk, including dependencies, auth, data, tests, accessibility, performance, and rollback.
Practical guideAGENTS.md vs CLAUDE.md: Rules Agents Can Actually UseChoose project instruction files for Codex, Claude Code, Cursor, and mixed-agent repositories, with a practical content template.
Ship and recover
Test the deployed path, preserve a known-good state, and make rollback part of completion.
A launch gate for AI-built apps covering accounts, data access, secrets, payments, errors, monitoring, SEO, and rollback.
Practical guideHow to Vibe Code a Website That Is Safe to LaunchAn end-to-end workflow for turning an idea into a working website, including auth, data, forms, testing, deployment, and rollback.
Continue by task