Deploys
The Deploys page: commits, CI runs, and post-deploy regressions connected to your scan history.
Most regressions ship. The Deploys page connects what shipped to what the scanner found, so a score drop comes with the commit list that preceded it instead of a shrug.
When the page appears
Deploys lives in the sidebar’s History group for projects that can supply deploy data: a linked folder with a git repository, or a GitHub connection. If the linked folder has no .git directory, the page says so and offers Add Folder; deploy tracking needs a real repository so SiteCMD can connect commits, scans, and regressions.
Reading your git history happens locally and read-only. Nothing about your repository leaves the machine.
What you see
Three stat cards up top:
- Total Commits, with the branch they’re counted on.
- Success Rate, computed from your GitHub Actions runs. It reads “Connect GitHub” until that integration is connected.
- Last Web Scan, the score and issue count of the most recent scan. Click it to open that scan.
Below them, Latest Commits lists your history in pages, each commit marked by whether it landed before or after your last scan. When commits have piled up since the last scan, a banner counts them and offers a one-click rescan, so “we shipped five things since we last checked” is never a surprise.
Post-deploy regressions
When the local correlation engine ties a score drop to a deploy, the page says so plainly: Deploy Likely Caused Regression when the confidence is high, Deploy May Explain Score Drop when it’s weaker. The alert shows the current score and open issue count, with two actions: Open dropped scan to read the evidence, and Scan after deploy to re-run against what actually shipped. The commit the correlation points at is flagged in the list.
The same deploy-linked score drops also raise alerts, so you hear about them even with the page closed.
GitHub
Connecting GitHub (you can start the device-flow connect from this page) adds two sections: CI / Actions, the recent workflow runs with result, branch, duration, and a link to the run, and Open Pull Requests. The Success Rate card fills in from completed runs. If syncing stops - an expired credential, usually - the page says “GitHub CI stopped syncing” and offers a reconnect instead of quietly showing stale data.
Recording deploys from CI
On a connected site, your pipeline can record deployments explicitly instead of leaving SiteCMD to infer them: sitecmd deploy records a deployment without code evidence, and sitecmd connected --submit records one with the checkout’s code findings attached. See the CLI reference.
AI editors get the forward-looking version: the MCP preview_deploy_risk tool takes the files a deploy is about to change and predicts which active issues are likely to regress. See AI editor setup.
Where else deploys show up
Every deploy becomes an event on the Activity timeline, alongside scans and incidents, and deploy context feeds the correlation notes you see on individual issues.