Guide 02 · OPTIONS

CONTRACT REVIEW · TEST_ONLYReview-candidate contract and local fixture evidence only · no distributed SDK, operational API, publisher parity, or decision authority

Compare options on a common basis.

Link each Option to its evidence, assumptions, scenarios, and trade-offs. Compare what supports or weakens it instead of assigning a recommendation score.

01

An Option is not an automatic ranking

An Option is a named alternative that people can compare. The current review contract defines no recommendation score, optimization result, or automatic-selection field.

  • SUPPORTS connects supporting material
  • CHALLENGES connects weakening material
  • MODELS connects a Scenario
  • SELECTS appears only on a human-recorded Decision
02

A Scenario needs an explicit horizon

A Scenario is a review frame under named assumptions, not a factual prediction. Fix fromAt and toAt, then reopen review when the horizon expires.

  1. 01
    State the baseline

    Separate what is supported now from what remains unknown.

  2. 02
    Vary assumptions explicitly

    Do not change several assumptions at once and hide the cause of a difference.

  3. 03
    Record each option’s trade-offs

    Keep constraints, sensitivity, and failure conditions beside benefits.

  4. 04
    Set the next review point

    New evidence or horizon expiry should create a ReviewTrigger.

03

Leave a structure for human review

  • Supporting and weakening evidence by option
  • Shared and option-specific assumptions
  • Horizon and sensitive conditions
  • Open EvidenceRequest records
  • Decision owner and review triggers