sh1-create_skill

Overview

This skill creates and manages batch workflow schedules with automatic dependency analysis and validation. It collects work items from GitHub, Jira, or local lists, detects dependencies in issue text, assigns priorities, and saves a deterministic schedule file. The schedule is stored as .wf/schedule.json and can be shown, edited, validated, or cleared from the command line. It is designed for straightforward, repeatable batch-run orchestration in shell environments.

How this skill works

The command fetches issues from configured sources (GitHub via gh, Jira via JQL, or local IDs), generates consistent work IDs, and scans bodies for dependency patterns (e.g., "depends on #N", "requires PROJ-N", "after: WORK-ID"). It builds a schedule JSON with works, priorities, dependencies, and progress, then checks for circular or unresolved dependencies and records the result. You can view the schedule, edit priorities or dependencies for specific works, validate integrity, or delete the schedule file with confirmation.

When to use it

• When you need a single, consistent batch schedule assembled from issues across GitHub, Jira, and local lists.
• Before running a batch workflow to ensure dependencies and priorities are valid and non-circular.
• To export or inspect a machine-readable schedule (.wf/schedule.json) for automation or reporting.
• When resolving dependency or priority conflicts across multiple tracked issues.
• To quickly rebuild schedules from updated labels, milestones, or JQL queries.

Best practices

• Keep dependency mentions consistent in issue bodies using the supported patterns for reliable detection.
• Run validate after creating or editing a schedule to catch cycles and unresolved references early.
• Configure priority rules in config.json to reflect your organizational batching strategy.
• Use source filters (labels, milestones, JQL) to limit schedule scope and avoid accidental inclusion.
• Back up .wf/schedule.json before running destructive edits or a clear operation.

Example use cases

• Create a nightly batch schedule from GitHub issues labeled 'scheduled' and save it for the runner.
• Aggregate current sprint work from Jira and local hotfix IDs into a single dependency-aware schedule.
• Show the current schedule to stakeholders including progress and per-work dependencies.
• Edit a work to raise its priority to unblock downstream tasks without changing the source issue.
• Validate a schedule before execution to detect circular dependencies and missing references.