Claude Code ships experimental advisor tool to consult a stronger model at key moments
anshulkundaje · x · 2026-09-25
Anthropic's Claude Code docs now include an experimental advisor tool (Anthropic API only) that lets your main model consult a typically stronger model at key moments — before committing to a plan, when stuck on a recurring error, or before declaring a task done. The advisor reads the full conversation and gives advice while the main model still writes the code.
The thread shares a practical setup: Opus 5.5 on high for the main session, three medium-tier subagents for reading code, editing/testing, and docs lookup, plus /advisor fable to put idle Fable quota on the sidelines as an advisor.
It also cites Anthropic's official Opus 5.5 guide: delete prompts like "think carefully" or "take a deep breath" from CLAUDE.md, since the model already decides how much to think — removing them speeds up responses with no quality drop.
More from coding & agent
- Meta launches Muse Code terminal coding agent, resets usage limits for DevDay — armand_ruiz · 2026-09-25
- User ditches GUI: local LLM now his main PC interface for installs, GRUB and more — Forward_Jackfruit813 · 2026-09-25
- Where an agent loop runs, how it fails, and how to build durable sessions — tempNull · 2026-09-25
- Community turns the Opus 5.5 pixel-art prompt into an open-source pixel-anims skill — majidmanzarpour · 2026-09-25
- Opus 5.5 + one prompt generates animated pixel art wizard purely in code — majidmanzarpour · 2026-09-25
- A visual harness for Claude Code: watching your agents work after hours — EricBuess · 2026-09-25