@jackkeepstabs@Replit Here's an example of how it would look for patching issues with unauthenticated endpoints.
After a separate subagent runs code review, you are shown a "Ready for review" pane so you can live test your app, or review the exact diff.
Changes merge back to main when you are ready.
@jackkeepstabs@Replit The agent applies fixes to security scan findings in a sandbox task that you can then review before applying the changes to the main branch for your project (very similar to a PR).
@Replit The migration task will:
- Provision your custom Clerk app
- Copy over existing users from Replit auth
- Update all application code and automatically test
Learn more here: https://t.co/zvivyxbHZ4
@Replit Until today, fully customized auth was only available to new Replit projects.
Now, you can easily move your existing projects over from Replit Auth. The migration happens in an isolated task, which lets you fully verify the change before applying it to main.
@elalvarobalbin@amasad We constantly monitor for newly reported vulnerabilities and notify builders + can even auto-open patches in builder's projects when a new critical CVE is reported: https://t.co/PvVT3wrSZ4
@Kaygeeartworks@amasad This is available to all paying customers on an opt-in basis for now. Our goal in the near future is to explore options where resolution of routine vulnerabilities can be fully managed by AI, even while you are away.
@Th3RealSocrates@amasad Auto-Protect is currently scanning for dependency vulnerabilities based on public advisory databases (for example, see https://t.co/4NfubbwVrt).
You can enable Auto-Protect only for vulnerabilities of a certain severity, such as high / critical only.
@PromptSlinger@amasad We recommend running Security Agent whenever you add a new major feature to your app, since it does a full teardown of your codebase.
We have other static analysis checks that are designed to run continuously as you build and in the background while you are away.