Automating the operations of a hundred repositories
I was managing many web applications, and the operations were manual: syncing project plans to issue trackers, checking fleet health, distributing code, scheduling newsletters, collecting data and scanning for security problems. It consumed hours every week across 4 CI/CD platforms.
What I built
Operations automation that manages more than 100 workflows across more than 100 repositories.
- A project sync agent that parses project documents, syncs them to GitHub Issues, detects schedule drift and projects completion dates.
- A fleet inspector that checks every repository for required files, commit freshness, open pull requests and plan status.
- A schedule analyzer that tracks velocity and flags overdue work.
- Hub-to-spoke pull requests that distribute shared file updates in a reviewable way.
- Daily and weekly email digests with engagement tracking.
- Semgrep, Nuclei and Gitleaks scans across the fleet every 12 hours.
- Tag-triggered publishing of 4 npm packages.
It runs on GitHub Actions, GitLab CI, AWS CodeBuild and Vercel.
What came of it
Fleet health is visible at a glance. Vulnerabilities are caught within 12 hours anywhere in the fleet. Newsletters go out on schedule. Packages publish when I push a git tag.
