AI Briefing
KO

CC 2.1.147: Code Review Agent Feature Update

·2026.05.25 01:47

Key point

The CC 2.1.147 update adds code review agent features supporting various difficulty and verification levels.

Details

A large number of new agent prompts for the /code-review command have been introduced. This update is designed to let users choose the precision and recall of code reviews.

Key Update Details:

  • Refined Effort Modes:

    • Low-effort: A mode that skips tests and fixtures to quickly find core errors.
    • Medium-effort: Focuses on Precision, returning up to 8 results.
    • High/Maximum-effort: Performs 5 independent analysis angles and multi-stage verification to maximize Recall, returning up to 15 results.
  • Advanced Verification Process:

    • A verification stage has been added that classifies review candidates into three levels: confirmed, plausible, and refuted.
    • Guidance has been included to treat uncertain candidates as 'plausible' unless the code explicitly refutes them, in order to increase recall.
  • GitHub Integration Feature:

    • A feature has been added that allows review results to be posted directly as inline comments on GitHub PRs via the --comment option.
  • Other Changes:

    • The 'Simplify' skill has been removed, and prompts related to automatic session name generation have been optimized.

This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.

Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.