run-execute

Deterministic run orchestration coordinates batch execution by initializing a run with a JSON array of work items, tracking state via YAML frontmatter, and updating run logs to ensure consistent artifact generation across all items.

Autopilot, confirm, and validate execution modes control run determinism by dictating whether work items process autonomously, require manual approval at specific stages, or need explicit validation before progressing.

Yes, YAML frontmatter manages execution state by tracking work item progress; the skill uses the 'yaml' library to parse and manipulate state files, updating statuses and intents as items move through planning, testing, and review.

Artifact synchronization occurs during the execution lifecycle by automatically generating plan.md, test-report.md, and review-report.md files while updating YAML frontmatter to reflect current item statuses and intents across the run.

Yes, the skill supports both single-item and batch wide execution, allowing you to initialize individual runs by specifying a single work item ID and intent, or coordinate wide batch runs containing multiple items simultaneously.