Compiler
Assembles the panel’s voices into the single answer that reaches the clarity check.
Algorithms route the question. Distinct human modes of judgement contest the answer. Select any personality to trace its public cognitive mapping and role in the council.


Retrieval can tolerate one register. Decisions that compound need competing disciplines: practical, strategic, ethical, sceptical and human.
Efficiency produces an answer. Judgement tests whether it should survive.
Each personality is a constrained reasoning role. The purpose is not theatrical disagreement. It is to make different forms of scrutiny visible before an answer reaches the user.
Asks what could break the proposed answer.
Asks what can actually be done under real constraints.
Asks who pays the price — reviewing only the question through a moral lens, never the panel's answer.
Assembles the panel’s voices into a single answer.
Reads the final answer cold and sends it back when it is not clear.
The personalities vary by question category. The sequence does not: Lead, Depth, Test, Ground and Human.
Classified before the panel is locked.
Clarifies what is actually being decided.
Judges context and hidden assumptions.
Challenges and stress-tests the answer.
Checks practical conditions and feasibility.
Checks impact, meaning and who is affected.
Compiled, audited and checked for clarity.
Assembles the panel’s voices into the single answer that reaches the clarity check.
Reviews only the question in parallel, never the panel's answer, and may also take a panel seat when ethics is central.
Sees only the final answer. If it is confusing, the panel must present its findings again.