Skip to main content
PR Blast Reviewer checks code changes for operational risks, including effects on dependencies and shared infrastructure. Read its findings in Anyshift, with optional publication to GitHub.
Availability depends on your project. Only project admins can configure or activate the reviewer. GitHub access and repository assignment must be ready before activation.

​
Quickstart

1

Connect GitHub

Connect the Anyshift GitHub App and assign the repositories to your Anyshift project. A GitHub organization owner may need to approve the installation.
2

Choose the repositories and review behavior

Open Agents → PR Blast Reviewer in the app, then Settings. Select the repositories to review and check Integration requirements for missing access.Choose whether results stay in Anyshift or are also published to GitHub. Configure automatic reviews, new-commit reviews and whether to include draft pull requests, then activate the reviewer when the requirements are met.
3

Read the review

In Activity, open a pull request to see its findings, supporting evidence and reviewed commit. Compare that commit with the current one: an earlier review does not assess changes pushed after it.

​
What the review covers

The review policy focuses on risks introduced by the change: dependencies and shared infrastructure, security, performance and resilience. Available production context determines what can be verified. A result can pass, need attention or remain inconclusive. Missing evidence or an incomplete review should not be read as a passed check. Review history keeps the verdict attached to the commit it assessed.

​
Where results appear

Choose Anyshift conversations only to keep results in Anyshift, or GitHub reviews and Anyshift conversations to publish review comments as well. Publishing requires the corresponding GitHub permissions. The reviewer does not push code, approve or merge pull requests. Merge enforcement remains in GitHub.

​
After the change ships

Use Deploy Watcher to check supported production changes after deployment. For an incident investigation, use Root Cause Analysis.