# Open Questions For CCRA

## QPI

1. Should `problem_owner` be required in QPI output for all calls, with `unknown` allowed when unavailable?
2. Should `time_scale` be required in QPI output, or remain a recommended routing field?
3. Is the current `dominant_scarcity` enum sufficient: `data`, `path_resource`, `consensus_order`, `mixed`, `unknown`?
4. For mixed cases, should QPI always output `classification=mixed`, or can it output a primary classification plus mixed evidence?
5. Are the violent reduction / malicious inflation checks strong enough for review, or should they become separate required output fields?

## Intellectual Archaeology

1. Should `recommended_max_depth` become a required model-card schema field?
2. Should `stop_reason` and `no_deeper_reason` become required output fields?
3. Should `value_of_deeper_analysis` be added explicitly, or is `no_deeper_reason` plus `action_implication` enough?
4. Should QPI approval be mandatory before Intellectual Archaeology can be selected, except for explicit model-extraction tasks?
5. Is `philosophical_bedrock` allowed only when the input explicitly concerns problem ontology or model foundations?

## Regression And Selector

1. Are 17 cases per model sufficient for the next review phase?
2. Should `regression_status` remain `pending`, or move to `in_progress` after this expansion?
3. Is selector v0.2's no-call threshold of `0.35` acceptable for review, or should it be calibrated with real inputs?
4. Is it acceptable for mixed QPI cases to recommend QPI first and leave IA as a later candidate only after QPI output?

## Status

1. Should `draft-callable` remain report-only for v0.2?
2. Should `draft-callable` become a first-class schema/status field later?
3. Should both models remain `draft / B / pending`? Current Codex recommendation: yes.
4. Should third model expansion remain blocked until after this review? Current Codex recommendation: yes.