Design review panel for an architecture proposal

The decision tree

5 typed decisions, each with an action for every answer

At fixed moments in a run, the engine puts one narrow question to a decision model. The decision model never writes text: it answers yes or no with a probability, picks from listed options, or gives a score, about a small slice of the material. The engine then does exactly what this tree says, which is what makes the run auditable. The thresholds are the template's design values, not measured results.

1. 1

   Before workers, before a task runs

   Small worker or strong modelYes or no, with a probability

   Before a task runs

   Does this design section describe storing, retaining, exporting or deleting data, so that the data reviewer must read it?

   Sees only: One design section and the data reviewer's brief

   Why: Each reviewer reads only the sections its standard speaks to.

   00.25 and 0.601

   • Yes: 0.60 or higher→thenAccept
   • Unsure: 0.25 up to 0.60→thenAccept
   • No: below 0.25→thenSkip this unit
2. 2

   Small worker or strong modelYes or no, with a probability

   Before a task runs

   Does this section involve authentication, secrets, network exposure or separation between tenants closely enough to need the strong security reviewer?

   Sees only: One design section and the security standard's table of contents

   Why: Spends the expensive reviewer on the sections where a missed gap is costly.

   00.30 and 0.601

   • Yes: 0.60 or higher→thenEscalate to a strong model
   • Unsure: 0.30 up to 0.60→thenEscalate to a strong model
   • No: below 0.30→thenAccept
3. 3

   After workers, the judge checks

   Evidence checkYes or no, with a probability

   After a worker answers

   Does the quoted design section omit or contradict what the quoted line of the standard requires, rather than cover it in other words in the same section?

   Sees only: The quoted design section and the quoted line of the standard

   Why: Separates real defects from a reviewer not recognising the author's wording.

   00.50 and 0.851

   • Yes: 0.85 or higher→thenAccept
   • Unsure: 0.50 up to 0.85→thenEscalate to a strong model
   • No: below 0.50→thenReject and retry
4. 4

   Reconciler, while merging

   Conflict checkYes or no, with a probability

   While reconciling

   Do these two reviewers' recommendations for the same section pull in opposite directions, such as longer log retention for debugging against shorter retention for privacy?

   Sees only: Two findings on the same design section

   Why: Conflicting advice is a decision for the board, so it is kept rather than averaged away.

   00.40 and 0.701

   • Yes: 0.70 or higher→thenMark unresolved
   • Unsure: 0.40 up to 0.70→thenMark unresolved
   • No: below 0.40→thenAccept
5. 5

   Accountable person, before anything is settled

   Person decidesYes or no, with a probability

   Before anything is reported as settled

   Would this finding, if upheld, require changing the design before build starts rather than noting a follow-up?

   Sees only: One verified finding with its quotes

   Why: The review board decides what blocks a build; the panel never approves a design.

   Accountable: The architect and the review board decide which findings block the build; the panel never approves a design.

   00.15 and 0.401

   • Yes: 0.40 or higher→thenAsk a person
   • Unsure: 0.15 up to 0.40→thenAsk a person
   • No: below 0.15→thenAccept