Anthropic's Claude Code now supports a read-only reviewer subagent, configurable via a markdown file. This subagent utilizes the Haiku 5.5 model, which is cost-effective at $0.10 per million input tokens for prompts up to 100K. The reviewer is designed to analyze code changes without the ability to modify files, ensuring an objective assessment. New features in Claude Code, such as an effort parameter and an autoCompactWindow setting, enhance the subagent's functionality and cost management. AI
IMPACT Enhances code review workflows by providing a dedicated, cost-effective, and objective subagent for analyzing code changes.
RANK_REASON This item describes a new feature/tool within an existing product, Claude Code, rather than a new model release from a frontier lab.
Read on dev.to — Claude Code tag →
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →