Dynamic Workflows — Vlad's Playbook
Opus 4.8's headline feature: Claude writes a script that plans a big task, fans out hundreds of parallel subagents, and verifies its own work before reporting back. What it is, how the generator→validator loop works, how to turn it on, where I point it, and when not to.
Dynamic Workflows — Vlad's Playbook Skip to main content New Edition 10.7 · Jun 18, 2026 New Chapter 47 — "The Measurement Layer": when the AI output IS the product, a three-line eval isn't enough. Anthropic's own course, ported to a TS eval harness + hybrid retriever (29 tests), now gating AI features before they ship. Read what's new → Edition 10.6 · Jun 16, 2026 New showcase — /good-taste: stop your AI from generating generic slop. A before/after of Leon Lin's public taste-skill (45k★, credited), with the receipts I built using it. Edition 10.5 · Jun 16, 2026 The design reading. Forty-five
saved by
related reading
- A Guide to Claude Code 2.0 and getting better at using coding agents – sankalp's blogsankalp.bearblog.dev
- Building Effective AI Agents \ Anthropicanthropic.com
- How I Use Claude Code | Philipp Spiessspiess.dev
- Best practices for Claude Code - Claude Code Docsanthropic.com
- Building Effective AI Agents \ Anthropicanthropic.com
- Shipping at Inference-Speed | Peter Steinbergersteipete.me
- How I use Claude Code on large projectslocalcan.com
- Don’t Outsource Your Thinkingteltam.github.io
- Arjun Virkarjunvirk.com
- Claude Opus 4.8: The System Card - by Zvi Mowshowitzthezvi.substack.com
- How Claude Code is built - by Gergely Orosznewsletter.pragmaticengineer.com
- Claude Code Cheat Sheetcc.storyfox.cz