What problem does it solve? When a project has self-hosted CI, developers waste time re-running tests locally that CI already passed, or incorrectly fall back to local runs when CI triggering fails. This Skill provides a deterministic decision flow for choosing where tests and lint execute. ## Core Features & Use Cases - CI Result Reuse: Checks existing GitHub Actions runs via gh run list and reuses successful results when the commit SHA matches the working tree HEAD, including baseline-branch checks for unpushed fix branches. - Strict No-Downgrade Rules: Distinguishes CI trigger failures (missing ref, auth errors) from CI unavailability, routing failures through AskUserQuestion instead of silently falling back to local runs. - Multi-Workflow Scheduling: Handles projects with separate workflows (e.g., macos-ci.yml for XCTest/lint, macos-xcuitest.yml for UI tests), triggering XCUITest only when UI behavior changes. - Use Case: Before pushing a bug fix on a new branch, the Skill detects the branch is unpushed, queries the base branch's CI run for the same commit, and reuses the passing result instead of running the full suite locally. ## Quick Start Ask the AI to decide whether to run the test suite on CI or locally for the current branch before pushing.
