Syntax

Dual-Layer Architecture

Human-Facing Form

Natural, semi-structured text. Written in the human's own language. Free-form but follows recommended patterns.

AI-Facing Form

Canonical, unambiguous structure with 5 sections:

Core Definition — Intent, Object, Constraints, Expected Output

Supporting Context — Context, Value, Actors, Evidence, Metrics

Development Layer — Conception, Vision, States, Transitions, Signals, Risks, Open Questions

Readiness Layer — Current Lifecycle State, Critical Gaps, Unknown Yet Items, Realization Decision

Linked Artifacts — References to external documents

Intent Vocabulary (8 Verbs)

Constraint Types (7 Types)

Evidence Markers

• [fact] — Verified information
• [to-collect] — Needs to be gathered
• [assumed] — AI or human assumption, needs confirmation

Signal Format

Signal: [type: severity] — description
  type = operational | quality | safety | feedback
  severity = info | warning | critical