Scope completeness is checked against backlog wording, linked docs, and likely implementation surface
Requirement gaps are flagged before the forecast range is trusted, instead of being hidden inside false precision
Dependency exposure, testing burden, and regression risk must be represented in the work plan
Recommended roles must match the stack, integration pattern, and rollout complexity
Estimate stability is blocked if ambiguity stays too high for a safe forecast
The output is a planning forecast for earlier pricing and resourcing conversations, not a delivery commitment
AI starts the scoping process, Dorsium improves it progressively, and supported language-aware agents can assist specific stacks and task categories when additional planning analysis is needed, with encrypted Supabase pgcrypto-backed storage behind the forecast trail.