Alibaba ships deterministic agent code review CLI
Open Code Review pairs hard constraints (file selection, rule matching, positioning modules) with LLM agents to fix incomplete coverage and position drift that plague general-purpose code review agents.
June 5, 2026
Summary
Eliminates surface-level diff feedback and chatty review quality swings by enforcing deterministic review scope and precise line-number accuracy. Developers get stable, complete changeset coverage instead of selectively reviewed code.
Why it matters
Eliminates surface-level diff feedback and chatty review quality swings by enforcing deterministic review scope and precise line-number accuracy. Developers get stable, complete changeset coverage instead of selectively reviewed code.
Implementation verdict
Replaces Claude Code with Skills and manual review workflows. Requires: LLM endpoint (Anthropic, compatible), Git repo, CLI install. Ready now—stable production code from Alibaba's internal use across tens of thousands of developers and millions of detected defects.
Sources
- 1.over the past two years, it has served tens of thousands of developers and identified millions of code defects
- 2.reads Git diffs, sends changed files to a configurable LLM via an agent with tool-use capabilities, and generates structured review comments with line-level precision
- 3.the root cause: a purely language-driven architecture lacks hard constraints on the review process
- 4.npm install -g @alibaba-group/open-code-review
Dev Signal
Get briefs like this in your inbox — free, 3x a week.
100+ sources compressed into one 4-minute read. Ranked, cited, implementation-ready.