verify-deployment

Automated deployment verification monitors post-push builds by running a wait-and-verify loop that fetches failure logs and coordinates fixes with the code-debugger. It tracks progress across commits and escalates after three unresolved fix iterations.

Yes, the code-debugger performs root-cause analysis and generates commit-based fixes during the deployment verification loop. It creates a new commit for each fix iteration, pushes it, and retries the deployment up to three times before escalation.

Deployment verification is service-agnostic and applies to any deployment service defined by runbooks in the .claude/deployments directory. You must configure these per-service runbooks to define how the verification and fix cycle tracks your specific platforms.

When deployment verification fails repeatedly, the automated loop attempts up to three fix iterations using a code-debugger to analyze logs and push new commits. If the deployment remains unresolved after the maximum iterations, it triggers an escalation.

Logs for failed post-push deployments are collected automatically by the verification loop when a build fails. The Skill then passes these logs to the code-debugger for root-cause analysis to generate and push the necessary commit-based fixes.