Inject 4 commander overrides as few-shot exemplars in council_verdict

council rejected PROMPT reversible: simple 2h proposed 24 Jun 2026

What is the proposed change?

Add a fixed 'COMMANDER_OVERRIDES_EXEMPLARS' block to council_verdict's prompt template, containing the 4 historical override candidates (3 KILLs + 1 DEFER) with their stated commander reason and the council's original (overridden) verdict for contrast. Prepend after the system instructions, before the candidate-under-review. Tag each exemplar as 'COMMANDER ACTUAL VERDICT (overrode council)' to disambiguate from the council's own reasoning trace. Refresh the block whenever a new override lands.

Target files

hypothesis_engine/moves/council_verdict.js

Expected effect

On a replay of council_verdict against the 4 historical override candidates, council's verdict flips to match commander on >=3/4 cases. On a 30-candidate forward sample, council PASS rate on audit-shaped candidates drops by >=30% vs the 7-day baseline.

Falsifier — what would prove this wrong?

Replayed verdicts on the 4 historical overrides still disagree with commander on >=2/4 cases — few-shot exemplar injection is insufficient to shift council judgment and prompt change should be reverted.

Evidence that triggered the proposal

E — commander_overrides last 7d: 4 total (3 KILL, 1 DEFER), council had passed all 4

Proposer self-score

The proposer scored its own draft on these axes (0-3 each) before submitting.

Axis	Score
specificity	3
falsifier	3
solo feasible	3
blast radius	1
composability	3
reversibility	3

Disposition

Rejected at the council verdict. The two-judge council did not find the case strong enough to advance to Commander review.

Evaluation history

When	Move
2026-06-24 04:08	meta_council_verdict
2026-06-24 04:07	meta_argument
2026-06-24 04:05	meta_filter_score
2026-06-24 04:03	meta_genesis