split-issue

To split a GitHub issue into tracked child work, you execute a previously-approved split plan that reads the parent issue's enriched spec and creates scoped child issues with correct labels, acceptance criteria, and parent-child linkages.

Automating GitHub issue splitting requires executing an approved split strategy that slices the parent's enriched spec into scoped child briefs, creating missing issues idempotently while updating the parent's split-children task list and evaluation mappings.

Yes, you need a pre-approved split plan. The workflow requires the parent issue to contain an evaluation block with a chosen split strategy specifying titles, acceptance criteria, files, dependencies, and model tiers for each child before creation begins.

Yes, the splitting process handles child issues idempotently by reusing existing open issues when possible, otherwise creating missing ones and recording the mapping back into the parent's evaluation block.

Issue splitting supports two test-ownership strategies: parent-owned integrated tests where the parent retains test responsibility, and sibling-owned integrated tests utilizing a synthetic test child issue for coverage.

Yes, newly created child issues can be auto-enriched by sequentially running an improve-issue process on them, adding further context and detail to the scoped child briefs generated during the split.