kiro-spec-batch

Spec batching automates end-to-end spec creation for roadmap features by dispatching parallel subagents in dependency waves. It initializes requirements, generates design, and creates tasks per feature with auto-approval, ensuring prerequisites complete in dependency order.

Dependency-wave orchestration groups roadmap features into execution waves so independent spec generation happens in parallel while respecting prerequisite dependencies. This ensures features with no dependencies generate first, followed by dependent features in subsequent waves.

Yes, the batch spec pipeline dispatches per-feature subagents that execute full spec pipelines in parallel. Each subagent initializes, generates requirements, produces design, and creates tasks for its assigned feature with auto-approval enabled.

Cross-spec consistency checks perform a final review after all waves complete, verifying that data models, interfaces, and naming conventions align across all generated specs. This ensures uniformity despite parallel generation by independent subagents.

The batch spec pipeline handles failures gracefully during wave execution. If a feature spec fails, the system manages the error without crashing the entire batch, allowing other independent features in the wave to continue processing.

Yes, roadmap-driven execution reads the roadmap to extract features, descriptions, and dependencies that drive the batch. A structured roadmap with dependency information is required to group features into deterministic execution waves.