LoginSign Up
← Back to all sparks
C

CodeRabbit

DEVOPS
Velocity8.8

AI-powered code review platform that posts line-by-line pull request feedback

coderabbit.ai ↗

CodeRabbit renames Autopilot to Autofix, staking the name on finishing PRs, not patching comments.

code-reviewpr-automationsecurity-scanningmerge-gatingcoding-agentsgitlab
◆Current state
CodeRabbit renamed two features: the one that fixes review comments is now Fix comments, and Autopilot, which carries a pull request to completion, is now Autofix. Underneath, the product has moved past comment-only review, with Triage rules that merge, close and fix PRs, a Security check that admins can require to block merges, Deep Scan findings available over the API on every plan, and a CLI that launches cloud Coding Agent tasks.
◆Where it's heading
CodeRabbit is repositioning from reviewer to agent that finishes the work, and the rename puts its main fix brand on end-to-end PR completion. Security review is turning into an enforcement layer wired into branch protection, while GitLab and Bitbucket parity and governance controls follow each new power the bot gains.
◆Prediction
Expect merge blocking and linked-repository scans to extend from GitHub to GitLab, following the parity pattern Triage and repository linking took over the past two weeks.

◆Recent moves

  1. 3d ago

    Autopilot renamed Autofix; old Autofix becomes Fix comments

    A naming change with no new capability: comment fixing becomes Fix comments, and Autopilot becomes Autofix. Giving the Autofix name to full PR completion shows where CodeRabbit wants its brand to sit, on the agent that finishes PRs.

    View source ↗
  2. 9d ago

    CLI v0.9.0 | CLI, Beta

    The CLI can now launch cloud Coding Agent tasks directly, with richer review JSON for scripting and signed Linux builds. It makes the terminal a front door to the same agent layer Triage uses, not just a local review runner.

    View source ↗
  3. 12d ago

    AI Deep Scan with linked repositories | GitHub Cloud | Essentials and above, AI Deep Scan access, Beta

    Deep Scan can pull dependency context from linked repositories, so a primary repo's findings account for code it consumes elsewhere. It builds on the automatic linking shipped earlier in the week and makes the scanner more useful in multi-repo organizations.

    View source ↗
  4. 12d ago

    Security merge blocking | GitHub | Advanced, Enterprise | CodeRabbit Cloud

    Admins can publish the Security check with a severity threshold and mark it required in GitHub, so unresolved findings above that bar block the merge. It turns security comments from advice into a gate, matching the broader shift from reviewer to enforcer.

    View source ↗
  5. 12d ago

    AI Deep Scan findings readable via API on every plan

    Deep Scan findings move from a Security add-on perk to something every plan can pull through a read-only API key. It extends last week's provenance work and makes security findings a data source for outside tooling rather than an upsell gate.

    View source ↗
  6. 12d ago

    Repository guideline files in the review summary | PR Reviews

    Review summaries now show which repository guideline files actually reached the reviewer and how each was found. It is a small transparency move that lets teams debug why a review did or did not apply their conventions, which matters more once reviews can trigger merges.

    View source ↗