Sergey Karayev's loop: Claude Code implements, a QA agent reviews, Claude iterates
sergeykarayev · x · 2026-11-03
AI educator Sergey Karayev shares his current coding workflow: Claude Code implements, an Astra-powered QA agent checks everything and suggests improvements, then Claude iterates.
The screenshot shows a real run: Claude Code reports adding v1 API endpoints for messaging ticket implementations and chats, with controller and policy tests passing (201 runs) — plus deliberately broken permission checks to confirm tests would fail. The QA agent's report covers feature works, lint, tests, added tests, and a failed code review, and Karayev asks Claude which suggestions it agrees with, prompting it to read the QA results JSON and keep working.
More from coding & agent
- Superserve (YC F26) launches agent cloud, already spinning up 100K+ machines weekly for YC and Lindy — ycombinator · 2026-11-03
- Lantern (YC) launches People Infrastructure for Agents, aggregating 70+ tools behind one CLI connection — ycombinator · 2026-11-03
- Rumored Claude Fable 5.5 beats Opus 5.5 and Gemini 4 Argon on three.js giraffe test — minchoi · 2026-11-03
- Dan Shipper: Partly Argon impresses while Antigravity feels stuck in 2025 — danshipper · 2026-11-03
- 'Opus 5.5 refactored my codebase into a Grosse Fuge — my team hates it, history won't' — eigenron · 2026-11-03
- Scale AI's READY framework: identical-accuracy agents need 39.2% vs 29.6% human review — demian_ai · 2026-11-03