sm:stage:chunking

Chunking breaks shaped stories, features, or design docs into flow-cycle-sized increments with explicit definitions of done. It creates bounded, testable pieces that AI agents can execute with clear boundaries and intentional handoffs.

Cost of delay sequencing orders chunks by economic, risk, flow, and joy costs to optimize delivery and learning. It prioritizes increments based on their impact on the overall workflow rather than simple chronological order.

Sequencing by cost of delay evaluates each chunk against economic, risk, flow, and joy costs. This scoring mechanism organizes flow-cycle-sized increments to maximize value delivery and learning during AI-assisted execution.

Yes, chunking applies to design-to-build handoffs across discovery, shaping, and execution phases. It transforms shaped narratives into bounded, testable chunks with explicit definitions of done for clear agent-based execution.

Yes, chunking requires a shaped narrative or design doc as input. The process transforms an existing story or feature into flow-cycle-sized increments, sequencing them by cost of delay to provide a ready-to-run starting prompt.

Chunking with guardrails and gates preserves agency, defines dependencies, and surfaces decision points before progress. It establishes clear boundaries for each increment to ensure intentional handoffs during execution.