Research report · data-report
Reviews code: a census of 110 listings
23 of 110 listings answer yes for “Reviews code”. Counted from our own datasheet, with the query that reproduces every figure.
On this page
Reviews code, across 110 settled listings
Of the 110 Published listings whose datasheet settles “Reviews code”, the largest group is “Does not apply” at 51 (46%). Measured 2026-09-09.
Explore the matching agents
Reviews code · Datasheet criterion reviews_code · Report counted .
Current listings as of ; this set may differ from the report's original sample. Includes “Yes” and “Qualified” answers only. Read the conditions on qualified answers before treating them as a match. “No”, “Not applicable” and “Not established” answers are excluded.
View 42 current matching records
- AgentBridge Qualified
It routes one agent's review of the other's diff into a running turn; the reviewing itself is done by Claude or Codex, not by AgentBridge.
Source retrieved 2026-09-02
- Agent Development Kit (ADK) Qualified
Reviewing a diff is an integration, not an ADK feature: the GitHub MCP toolset lets an agent read and act on pull requests.
Source retrieved 2026-09-02
- Agentix Labs Qualified
It reviews an existing deployment's configuration, secrets and connector boundaries; nothing on the site offers pull-request or diff review.
Source retrieved 2026-09-09
- AgentsKit.js Yes
Two review personas ship in the skills package and pair with the GitHub tool, returning approve, request-changes or comment verdicts on a diff someone else wrote.
Source retrieved 2026-09-09
- Agno Yes
Code review over a diff is one of the documented coding-agent patterns.
- Aider Qualified
Documented as reviewing a diff or PR branch you feed it via /run git diff and /ask; there is no automated review of someone else's pull request.
Source retrieved 2026-09-02
- Augment Code Yes
Deep Reviewer, Pair Reviewer and Risk Analyzer experts review pull requests and post inline findings.
Source retrieved 2026-09-02
- AutoGen Qualified
The reflection design pattern pairs a coder agent with a reviewer agent that critiques the generated code; no surface reviews a diff or pull request someone else wrote.
Source retrieved 2026-09-02
- Bolt Qualified
Audits your own project's code and database on paid plans; it does not review a diff someone else wrote.
Source retrieved 2026-09-02
- Claude Agent SDK Yes
The CLI surface is documented as a reviewer over a diff piped in from git or gh.
Source retrieved 2026-09-02
- Claude Code Yes
The /code-review command reviews a local diff on any plan; the managed GitHub PR reviewer is a research preview on Team and Enterprise.
Source retrieved 2026-09-02
- Cline Yes
Reviews diffs against rules, and the docs ship a GitHub Actions workflow that comments on PRs.
Source retrieved 2026-09-02
- CodeRabbit Yes
Reviewing someone else's diff is the core product: automatic line-by-line review on every pull request.
Source retrieved 2026-09-02
- Codex CLI Yes
/review runs a dedicated reviewer over a diff, commit or branch without editing the tree.
Source retrieved 2026-09-02
- Cursor Yes
Two review paths: Bugbot reviews pull requests in the repo host, and Agent Review reviews local changes before commit.
Source retrieved 2026-09-02
- Devin Yes
Devin Review reviews pull requests anyone authored, reorganising and explaining large diffs in the Devin webapp.
Source retrieved 2026-09-02
- Devin Desktop Yes
Quick Review puts a second agent over the diff; the SWE-check model is free on every tier, GPT 5.5 and Opus 4.7 bill per token, and Enterprise admins must enable it. Devin Local agent only.
Source retrieved 2026-09-03
- Factory Yes
/review reviews branches, commits or uncommitted diffs locally; a CI workflow reviews pull requests.
Source retrieved 2026-09-02
- Gemini Code Assist Yes
A separate GitHub app reviews and summarises pull requests and answers follow-up questions via /gemini comments; files under .github/workflows are excluded.
Source retrieved 2026-09-02
- Genspark Qualified
Claw can review pull requests through a saved GitHub login (listed under Developer Tools in its Services tab); no diff-review surface exists in Genspark Code itself.
Source retrieved 2026-09-02
- GitHub Copilot Qualified
Reviews pull requests on GitHub and file diffs in editors, with suggested fixes; the comparison table marks both rows Not included on Free.
Source retrieved 2026-09-02
- Glean Yes
The pull request review agent fetches a PR diff, checks it against style guides and posts comments back to GitHub.
Source retrieved 2026-09-02
- Google Antigravity CLI Qualified
The diff viewer reads any VCS, turn or commit diff and forwards your line comments to the agent; no agent-authored review of someone else's diff is documented.
Source retrieved 2026-09-02
- goose Yes
A first-class CLI verb reviews a diff you did not write, and it picks up repo-local review checks from .agents/checks/*.md.
Source retrieved 2026-09-02
- HOL Guard Qualified
The sibling plugin-scanner reviews repositories and gates pull requests on security grounds — prompt-injection markers, dangerous shell commands, secret leaks — rather than reviewing a diff for correctness like a code reviewer would.
Source retrieved 2026-09-07
- Junie Yes
Dedicated read-only review subagent for local diffs (/review) and for opened or updated GitHub pull requests, with inline comments.
Source retrieved 2026-09-02
- Kilo Code Yes
Kilo Code Reviewer analyses pull requests and comments in the PR.
Source retrieved 2026-09-02
- Letta Yes
Reviews diffs on demand or on every PR via the GitHub action's prompt input.
Source retrieved 2026-09-02
- Moveworks Qualified
It reviews agent builds it helped assemble, not a diff written by another engineer.
- Muse Code Qualified
Vendor claims background reviewer agents, and peer sessions can be sent review requests; no flow for reviewing an external diff or pull request is documented.
Source retrieved 2026-09-02
- OpenAI Agents SDK (Python) Qualified
File-grounded review and per-reviewer sandbox workspaces are documented use cases; no diff or pull-request review surface is described.
Source retrieved 2026-09-02
- OpenClaw Qualified
No built-in review product; the vendor's showcase describes a community-built loop in which OpenClaw reads a PR diff and answers with suggestions and a merge verdict.
- OpenCode Yes
Invoked on a review comment it receives the file path, line numbers and diff context, and PR-open events can trigger it automatically.
Source retrieved 2026-09-02
- OpenHands Yes
A GitHub Actions PR-review agent posts inline comments with priority labels on someone else's diff.
Source retrieved 2026-09-02
- Perplexity Qualified
Review happens through the managed GitHub connector, which the same note marks as preview; there is no dedicated diff-review product.
- Pydantic AI Qualified
Review comes from the third-party Macroscope CLI, which you must install and authenticate yourself; the capability surfaces findings only and the agent fixes them with its own tools.
Source retrieved 2026-09-02
- Salesforce Agentforce Qualified
Static analysis with auto-fix runs over code in your own workspace and Git diffs can be attached as context, but no diff or pull-request review feature is documented and the docs state the trust layers do not replace code review.
Source retrieved 2026-09-03
- Sim Yes
Review Code mode checks out an existing PR read-only and submits a GitHub review with inline comments.
- Vecbase Yes
Reviews a PR link or diff and returns findings with file and line references, severity and suggested patches.
Source retrieved 2026-09-03
- Warp Yes
Cloud agents review other people's pull requests and comment inline, via GitHub Actions or the GitHub App; the in-app review panel is the inverse, a human reviewing the agent.
Source retrieved 2026-09-03
- Zed Qualified
A branch diff can be @-mentioned into a thread as context, but Zed documents no pull-request or third-party-diff review surface; its review UI covers the agent's own edits.
Source retrieved 2026-09-03
- zot Qualified
Review is not in the binary: it comes from a separately installed extension by zot's own author, listed in the vendor's extension directory.
Source retrieved 2026-09-03
The population
110 listings in this index are Published. This is the population as of 2026-09-09. 110 of the 110 published listings in this index carry a settled answer for “Reviews code”. That row asks: Does it review a diff someone else wrote? Every listing in the index is settled on this row, so nothing is left out.
Findings
-
23 of the 110 settled listings answer “Yes” for “Reviews code”. That is 21% of the settled set.
-
19 of the 110 settled listings answer “Qualified” for “Reviews code”. That is 17% of the settled set.
-
11 of the 110 settled listings answer “No” for “Reviews code”. That is 10% of the settled set.
-
51 of the 110 settled listings answer “Does not apply” for “Reviews code”. That is 46% of the settled set.
-
6 of the 110 settled listings answer “Not established” for “Reviews code”. That is 5% of the settled set.
1 of those yes answers is Codex CLI, which still answers yes for “Reviews code”. Its listing carries the stored answer and links to the source recorded for it. This is one worked example, not an independent audit of every source in the census.
What we counted, and how
Each figure above is a count over the “Reviews code” row of the listing datasheet, taken from the same stored answer the listing page renders. The denominator is the 110 listings whose answer is settled, meaning one of yes, qualified, no, does not apply, not established. On this row that is the whole index, because every listing has a recognised stored answer. Every number here is stored with the read-only query that reproduces it and re-run every sixty seconds against the live corpus, so a figure that stops reproducing surfaces as drift rather than as a stale sentence nobody notices. Our full method covers how a datasheet row is settled in the first place.
Limitations
This counts stored datasheet answers, not independently tested capabilities. A sourced answer can record a vendor statement or our reading of published evidence; this census does not re-fetch those sources. “Not established” means we have not established an answer. That can reflect vendor nondisclosure, blocked evidence, or unfinished research, not a no. An absent or unrecognised answer is excluded rather than treated as a researched finding. The figures are restated when the stored corpus changes; the date above is the count used for this published version, not a new verification of the vendors.
Get the next report
New agents rankings and fresh data reports. One short email, one-click unsubscribe.