Interrogating Design
Interview me about every aspect of this decision until we reach shared understanding. Walk down each branch of the decision tree, resolving dependencies one by one. For each question, provide your recommended answer. Ask one question at a time. If a question can be answered by exploring existing documents, explore them instead.
OODA Loop Mode (Boyd)
Run the interrogation as an iterative cycle:
- Observe: Read all relevant files, grep for related terms
- Orient: Synthesize against foundational constraints — destroy outdated mental models if new data contradicts them
- Decide: Formulate a specific hypothesis
- Act: Test the hypothesis against the repo
Dependency-Order Walkthrough
Check implications in layer order (customize for your universe):
- Foundations — contradicts core axiom or definitions?
- Physics — violates particle mechanics, fields, or temporal rules?
- Chemistry/Biology — contradicts established life sciences?
- Civilizations & Narrative — contradicts existing canon?
Adversarial Red Teaming Escalation
- Does it contradict any explicit rule?
- Does it contradict any implicit constraint (downstream implications)?
- Does it survive cross-civilization application?
- Would a hostile reader find it inconsistent?
- Strip the narrative beauty — is the bare mechanism still valid?
Russell Conjugation Check
Restate the proposal in the most boring, clinical language possible. If it only works in emotive framing, it's aesthetic, not physics.
After Validation
- Self-review for placeholders, contradictions, ambiguity
- Present in digestible sections
- Save to
meta/design-decisions/YYYY-MM-DD-<topic>.md - Transition to /planning-worldbuilding or /writing-science
Out of Scope
Does NOT write content. Does NOT modify files. Only interrogates and validates.