Know the risk before you merge

Run /merge-risk on any pull request to evaluate alignment, impact, verification, rollback cost, and error detection.

Try Warp FactoriesView on GitHub

Five signals behind every merge-risk grade

/merge-risk scores across 5 criteria: alignment, impact, verification, rollback cost, and error detection. These are weighted to find the final merge risk score, which you can use to focus human attention on the most important changes.

  1. Alignment

    Keeps the change tied to the issue and PR description, surfaces undocumented assumptions, and calls out drive-by edits.

  2. Impact

    Finds the blast radius and flags work in sensitive paths: auth, permissions, payments, PII, migrations, dependencies, and infrastructure.

  3. Verification

    Checks whether tests cover the behavior and whether user-facing changes include computer use screenshots or recordings that prove the experience.

    Enable computer use in your factory

  4. Rollback cost

    Separates reversible two-way doors from destructive or hard-to-undo changes, so reviewers know how to recover quickly.

  5. Error detection

    Checks whether logs, monitoring, and alerts expose failures before users report them.

    Connect production signals to your factory with webhooks

Computer use catches what tests miss

Give every user-facing change end-to-end proof. Enable computer use in your factory

“The moment that convinced us was watching a QA agent run a real screen reader and fail its own team’s PR because a required announcement was missing.”

— Warp Customer

Use Warp Factories to automate the whole SDLC

Defined in code, built for scale, and easy to deploy.

Request early access

Set up your first factory with early access.