01
- Rescue
- Builds stuck red, deploys failing, caches missing. We get the pipeline passing again and only then write down how to keep it that way.
02
- Cost and minute review
- Which workflows are eating the plan allowance, which could run on cheaper runners, which could be skipped by path filters. Numbers before and after.
03
- Security hardening
- Pinning actions to SHAs, scoping GITHUB_TOKEN permissions, moving cloud deploys to OIDC, auditing org and repo secrets so rotations actually happen.
04
- Self-hosted runners
- Standing up ARC on Kubernetes, ephemeral runners that reset between jobs, and a fleet that can be observed rather than hoped for.
05
- Reusable workflows
- Common build, test and deploy jobs lifted into a shared workflow repo so teams are not each rewriting the same YAML slightly differently.
06
- Managed Tech
- Workflow upkeep across your org, deprecations handled before they bite, minutes tracked monthly, and someone to call when a deploy workflow hangs on a Friday evening.