# Publication review: ClearFold Laundry

The author used Codex to check this example before publication. This is publication quality control, not a second council verdict and not a required stage for users. The council itself ran entirely in Claude Code with the recorded Sonnet 5.5 model.

The author subsequently confirmed medium reasoning for this run. This is author-supplied configuration context, not independent runtime verification. Future model or setting changes may produce different results; this run is not a comparison.

## What was checked

The brief matches the separately retained authored original and its SHA-256. All 12 saved packets contain that brief. The five reviewer packets contain complete anonymized answers in ABCDE, BCDEA, CDEAB, DEABC and EABCD order. The replacement chair packet contains all original advisor answers and reviews. These checks verify consistency within the supplied record, not extraction from underlying host logs or runtime isolation.

Independent calculations prepared before the run confirm 13-week incremental earnings of $10,000 for A and $9,565 for B; minimum cash of $11,550 and $18,045 with agreed payments; and $8,200 and $17,500 with two-week delays. B is only a proposal. A's second 13 weeks are not forecast. This is not evidence of a real business outcome, superiority to a direct answer, or a model comparison.

## Qualifications

- Four reviewers criticized the Outsider partly for not providing a recommendation or calculations despite its different assignment. Reviewer 5 defended that purpose. The chair's claim of unanimity is incorrect.
- Some claimed omissions were already covered: the Executor proposed B plus a conditional A in one message before Reviewer 3 said nobody had. Peer review corrected some numbers but did not consistently check all assertions.
- The first chair packet inserted five words into an advisor answer. It was superseded by a corrected complete packet. Recovery was disclosed, but comparison occurred after dispatch.
- Three advisors, not four, exceeded 500 whitespace-delimited words (518, 581 and 517). All reviewers exceeded 200. The transcript's summary of advisor lens lines mistakenly repeats output instructions, and the final chair word-count label contains an unexpanded expression; the full packets and returned text remain available.
- $1,800 is the modeled cash shortfall under exactly a two-week delay, not a fully specified deposit contract. Receipt date, invoice credits and repayment terms affect subsequent cash; deposits must not be counted as extra earnings.
- The original audit's concluding statement that a “≥2-week payment term” covers four weeks of delay is wrong. Two weeks plus four weeks of delay creates the six-week collection lag that breaches the floor. Its earlier statement that two-week-or-faster terms pass a two-week delay is consistent with the model.
- The audit describes baseline weekly surplus of $1,000/$1,900 as before fixed costs. Those figures are after the $4,400 fixed costs.
- Stress-case failure supports a cautious recommendation but does not establish a likelihood of late payment. Requiring survival of that stress is a conservative decision criterion. The brief already stipulates confirmed overtime; asking for written confirmation is an implementation safeguard, not discovery that capacity was unavailable.

## Presentation

Feature one business example. Label condensed prose as summaries of the council output and keep the original analyses available. Put the small-business decision before execution details. State that one supported harness can run the entire method with one model; portability means choosing an environment, not combining subscriptions. Keep this review optional and preserve material conditions beside the recommendation.