CI/CD

Trace the affected event through workflow creation, job selection, dependencies, commands, artifacts, and deployment. Diagnose the first failing transition from its configuration and run evidence.

Read provider behavior, then the affected provider: GitHub Actions, GitLab CI, or [Bitbucket Pipelines][ref-1].

Use provider-native YAML. Keep required failures visible. Pin executable dependencies and identify artifact producer revisions. Trace control of checkout content, interpolated inputs, caches, artifacts, runners, and credentials before privileged execution.

Validate changed syntax, event selection, dependency flow, and commands. Use hosted run evidence for trigger, permission, runner, and approval behavior. A pipeline repair must not bypass a failing gate or deploy an unverified artifact.