speckit-tasks

You can generate a dependency-ordered task list by feeding design artifacts like plan.md, spec.md, and data-model.md into a task generator. It translates these inputs into a structured tasks.md file, grouping actionable work items by user stories in priority order.

The best way to structure a tasks.md file is using a phase-based approach: Phase 1 Setup, Phase 2 Foundational, Phase 3+ User Story phases, and a Final Polish phase. This maps tasks to user stories like US1 and US2 with clear dependency ordering.

Yes, you can parallelize tasks when planning from a spec and data model. The generated task plan includes parallelization hints and a dependency graph, indicating which actionable work items across different user stories can be executed concurrently.

You do not strictly need a contract file, but including a contracts directory improves the generated task plan. The generator converts plan.md, spec.md, data-model.md, and contracts into a unified, self-contained checklist mapped to user stories.

Mapping tasks to user stories improves project delivery by enabling independent validation and tracking. It organizes the dependency-ordered task plan around specific user needs like US1 and US2, ensuring each phase delivers testable, scoped functionality.

A limitation of auto-generating a task plan from markdown specs is that the output quality depends entirely on the input artifacts. If the plan.md, spec.md, or data-model.md lack detail, the resulting dependency-ordered checklist may require manual refinement.