What problem does it solve? AI coding assistants often skip tests, rubber-stamp their own verification, and lose track of requirements on long runs. QRALPH wraps Claude Code in a deterministic 14-phase state machine that enforces every step of software delivery — requirements decomposition, TDD, quality gates, fresh-context verification, and post-deploy smoke tests — so nothing gets silently dropped or bulk-approved. ## Core Features & Use Cases - Deterministic 14-Phase Pipeline: IDEATE through COMPLETE, with the Python pipeline script deciding agent selection, model assignment, parallel execution groups, and gate enforcement while Claude does the creative work. - Evidence-Based Verification: A fresh-context verifier checks every acceptance criterion against actual source files with file:line quoted evidence, and quality gates (lint, typecheck, tests) hard-block progression on failure. - Full Lifecycle Automation: Persona generation, concept review, parallel execution with early-start optimization, quality loops with backtracking, demo gates, git branch/PR automation, deployment, smoke tests against the live URL, and cross-project learning capture. - Use Case: Run QRALPH "Add user profile page with avatar upload and privacy settings" and the pipeline decomposes requirements into tracked fragments, spawns parallel implementation agents, runs quality loops until convergence, verifies every criterion independently, and optionally deploys with smoke tests. ## Quick Start Ask the AI to run QRALPH with your project request, for example: run the qralph pipeline to build a landing page with a contact form and deploy it to Cloudflare Workers.
