the-jury

Ask to convene a jury on your decision. The skill spawns 3 to 5 subagent jurors with distinct personas and critical lenses, collects blind independent votes, runs anonymized deliberation, and returns one committed verdict with a confidence level.

It fits contested, high-stakes choices such as engineering tradeoffs, architecture migrations, build-vs-buy, vendor selection, hiring, and research design. It is not meant for simple factual lookups, pure critique without a decision, or writing the solution itself.

Round 1 votes are blind so no juror sees others first, deliberation is anonymized, and every flip must cite a specific new argument. If flips look like majority-chasing, the tally falls back to the independent Round 1 aggregate at low confidence.

Yes. Without parallel subagents, the panel is simulated as sequential role-played passes in one context, but all Round 1 positions must be generated before any position is revealed to any juror, preserving the blindness requirement.

The foreman never abstains or coin-flips. Ties are broken on the merits: the option with the higher evidence grade that best survived the devil's advocate's strongest challenge wins, and the closeness is recorded as dissent with lowered confidence.

Skip it for simple factual lookups, tasks where you only want critique without a decision, or when you need the agent to build a plan or write the solution. It also adds cost and latency, so trivial decisions do not justify a panel.