AI-Augmented Development Workflow

AI-Powered

⬢Developer Workflow

My daily development workflow built around Claude Code as the central AI pair programmer. Linear for issue tracking and sprint planning, GitHub CLI for branch management and PR creation, Playwright MCP for live browser testing during development, and Vercel for instant preview and production deploys. Every project — from Cassandra Furs to Eventhor — ships through this pipeline.

Technology Stack

Claude CodeLinearPlaywrightGitHub CLIVercelFirestoreVS Code

System Architecture

Loading diagram...

Specifications

▸Claude Code: AI pair programming — codebase navigation, multi-file edits, test writing, debugging via conversation
▸Linear: Issue tracking with priority triage, sprint cycles, and project roadmaps
▸Playwright MCP: Live browser automation during development — visual verification without leaving the terminal
▸GitHub CLI: Branch creation, PR workflows, code review, and merge — all from the command line
▸Vercel Deploys: Every PR gets a preview URL, merge to main triggers production deploy
▸Firestore: Real-time database for dynamic content — products, user data, configuration

Design Decisions

Challenge	Solution
Context switching between tools	Claude Code as the hub — reads code, runs commands, creates PRs, all in one session
Manual browser testing is slow	Playwright MCP lets Claude Code interact with the live site directly
PR descriptions are tedious	Claude Code generates PR title and summary from commit diff analysis
Need visual QA before merge	Vercel preview deploys on every PR — check the live URL before merging
Data changes shouldn't require deploys	Firestore for dynamic content — update DB, site reflects changes instantly

Process Flow

Loading diagram...