planning

Converting a PRD into testable tasks involves decomposing feature specifications into discrete actions with acceptance criteria. This process generates a structured plan where each task is uniquely identified, testable, and traceable to specific design decisions within your documentation.

Generating a dependency graph from a technical design document requires mapping task relationships during decomposition. The process outputs a structured file like tasks.json, applying DAG validation to prevent cycles and ensuring accurate dependency tracing across project phases.

Ensuring traceability from an ADR to development tasks requires mapping each decomposed task back to specific architectural decisions. This creates a verifiable link from your design documents to executable actions, guaranteeing that every task aligns with established project constraints.

Yes, validating cycles in a task dependency graph is a core function of this workflow. It applies DAG validation checks to your generated task list, preventing circular dependencies and ensuring your project phases execute in a logical, unblocked sequence.

You need a worktree to execute the planning workflow and output your tasks.json file. Running the process in a worktree ensures the generated task list, dependencies, and traces are committed directly to your repository before implementation begins.

Tasks.json serves as a unified output for orchestrating software project implementation. It captures decomposed feature specs with unique IDs, dependency graphs, and design traces, providing a single file to guide single-task-per-commit delivery and validation gates.