Council

Caller challenge

One consensus shape is never synthesized: the judges agree the caller's stated direction is wrong. Independent agreement against the caller is a strong signal, and it is still not authority: the caller holds context no judge was given, and a synthesis that folds the judges' position into a recommendation deletes that context without telling anyone it was overruled.

When judgments recommend a change to something the caller specified (merging what they separated, cutting what they asked for, reversing a declared direction), record it as a caller_challenge entry, not a consensus point. Use these five fields; optional judge_count requires at least two supporters, while disagreement_kind classifies the objection:

• caller_stated: their direction, in their words, not paraphrased.
• judges_recommend: the change, who supports it, and whether their views were independent or peer-informed; never describe debate votes as independent.
• reasoning: the case at its strongest.
• context_possibly_missing: what the judges provably were not given. This is the field that makes the entry honest and the one most likely to be dropped; an entry without it is majority laundering wearing a new label.
• cost_if_wrong: what breaks if the caller's direction was right.

The caller's direction is the report's default and stays the default; the burden of argument is on the judges. When the judges classify the change as a security or feasibility defect rather than a preference, say which (disagreement_kind); the caller still decides, knowing the kind of disagreement.

The named failure mode is quiet adoption: a council that converges against the caller and returns a synthesis reading as if the caller had asked for the judges' version all along. Stop condition: every judgment that contradicts a caller-stated direction appears in caller_challenge with all five fields, or it does not appear in the report at all. Whether the challenged decision can be undone belongs in Plan, with actual undo cost and existing authority; the council must not assume either.

Synthesis section

The report ends with an explicit consensus/divergence synthesis: consensus points with their methodology spread, divergence points with each side's cited evidence, minority findings preserved in their own words, unresolved assumptions, and any caller_challenge entries. Synthesis is complete when every judge finding lands in exactly one of those buckets; a finding silently dropped from synthesis is majority laundering.

Output

• Destination: caller-selected protected external non-Git storage; preserve existing legacy evidence. Missing routing is not a workspace fallback: when no destination is supplied, return the report inline in the conversation and write no file.
• Filename: council-report.json.
• Format: council-report.v1 JSON: the frozen question and subject digest, every judge's context ID, evidence methodology, cited evidence, and disclosed omissions, plus the consensus/divergence/minority/unresolved synthesis and any caller_challenge entries. Record mode and candidate digest in each judgment and methodology and source references in their existing fields; no new schema is needed. It carries no verdict, readiness, or PASS field; the validator rejects one.
• Validation command: this skill's scripts/validate-output.sh <council-report.json>.

A judge that times out, errors, or returns an evidence-free judgment is excluded from agreement counting and recorded as non-returning; if fewer than two eligible initial judgments remain, report insufficient independent coverage rather than synthesize a thin consensus. If no valid report can be formed, return the incomplete outcome and available receipts without fabricating judge records.

Prompt

Use /agentops:council to compare architectures for reliable Job redelivery.
Use four distinct available models I authorize for this source. Have each
propose an approach independently, then debate the alternatives. Require
three of four to support the same exact recommendation. Cap debate at five
rounds and the whole council at 60 minutes. Preserve objections and explain
what evidence we still need before implementation or validation.

Resolve the actual authorized model pins before dispatch; these example bounds are caller choices, not skill defaults. For validation, provide the unchanged acceptance and exact candidate, and return findings to the fresh validator without voting on PASS.

It's working if

• Initial views are sealed before cross-review; any debate is bounded and labeled peer-informed, with exact-candidate votes and dissent preserved.
• Every judge finding lands in exactly one synthesis bucket; none is dropped.
• A judgment that contradicts a caller-stated direction appears as a caller_challenge entry with all five fields, never as a consensus point.
• Every consensus claim names at least two distinct evidence methodologies, or is labelled single-method agreement and weighted as one confirmation.
• No verdict, readiness, or PASS field appears anywhere in the report.

Boundary

Council does not mint a verdict of any version — no PASS/FAIL/NOT_PROVEN, no verdict.v* — edit the subject, retry work, choose a next action, or authorize Git, closure, release, or delivery. When Council is used as a Validate strategy, one accountable fresh validator consumes its report and Validate remains the sole semantic result owner and the only optional verdict.v2 writer.