What problem does it solve? Polling for GitHub Actions completion often fails silently: checking only whether the latest run is completed can match an old, already-finished run, causing the process to exit immediately with stale results. This Skill waits correctly by distinguishing newly triggered runs from previous ones. ## Core Features & Use Cases - Run ID-based detection: Captures the previous run ID before triggering a workflow, then waits until a different run reaches completed status, avoiding false positives from old runs. - Event-aware filtering: Filters by workflow event (push, pull_request, issue_comment, workflow_dispatch) instead of branch, since slash-command-triggered runs report the base branch as headBranch. - Background waiting with result interpretation: Runs a single background polling loop and interprets the conclusion (success, failure, cancelled, skipped), classifying failures as code-related, approval-gated, or infrastructure issues. - Use Case: After pushing to a PR or triggering a workflow via a PR slash command, invoke this Skill to block until the new CI run finishes, then report the run ID, conclusion, and failure cause classification. ## Quick Start Wait for the GitHub Actions workflow I just triggered on this PR to finish and tell me the result and why it failed if it did.
