Claude Code tip: run 5 subagents to audit 20 things before shipping a vibe-coded site
udmrzn · x · 2026-09-28
A practical Claude Code workflow: before shipping a vibe-coded site, have Opus 5.5 dispatch 5 subagents to audit 20 items and fix them in one pass. Highlights:
- Design consistency: write a DESIGN.md first; constrain colors, font sizes, spacing and radii to it; keep type hierarchy clear; make same-kind buttons/cards/inputs look alike; ensure text is readable on dark and light backgrounds
- Mobile: no horizontal scrolling or overflow, add a mobile menu, big-enough tap targets, layout survives text scaling
- States: proper loading/empty/error screens, hover/pressed/disabled feedback on every button, inline form errors plus submitting/success/failure messages
The post also cites Anthropic's new full cost guide for Opus 5.5: at the same effort level it thinks more per turn than Opus 5 (especially at xhigh/max), thinking is billed as output, and output tokens cost 100x a cache read — so start at medium and reserve xhigh/max for measured gains. Includes a prompt for Opus 5.5 to review your Claude Code setup (settings, subagents, CLAUDE.md, MCP servers) against the guide.
More from coding & agent
- JevControl open-sources a tool to test if smaller models can run your agent's routine choices — airesearch12 · 2026-09-28
- Claude Code CLI connects to Gmail easier than Google's own Antigravity, dev finds — randal_olson · 2026-09-28
- Pelican test on MiMo 2.6: plan mode cuts agent calls to 12-17 vs heavy iteration — PilgrimofHaqq2 · 2026-09-28
- Memory history is not reasoning history: agents lack a causal provenance layer — puppy_lover_2021 · 2026-09-28
- mark_k: AI coding is approaching a step change — systems only AI can debug — mark_k · 2026-09-28
- Livestream: Sentinel, a Real-Time Node Graph Built Natively for Agents — PurzBeats · 2026-09-28