# Review status of the evidence resources

Version 0.3.2 is a fixed, LLM-assisted literature extraction prepared for reuse. Author verification is incomplete. It must not be described as a fully author-verified dataset, an independently double-screened review, a validated clinical standard or an executable benchmark.

The author withdrew the fixed 70-record sample-check plan on 8 November 2026. It was not performed. The original sample (40 title/abstract and 30 full-text decisions) is retained as a historical record and is no longer a required task. The optional OpenAlex candidate list and the blank 85-row review form do not document completed author verification.

Authors explicitly decided borderline inclusion and clinical-setting rules. ASD is excluded as a general skill-generation example using nutrition calculations. Coach Iris is coded by the retrospective design of its FAST assessment, while its operating patient service is documented separately. These author decisions do not certify the remaining per-study extraction or screening decisions.

Assistant source checks and deterministic consistency checks have been performed and are documented in the project records. They do not substitute for author review. Corrections will receive a new version; the paper-linked snapshot will remain recoverable.

No new patient data were collected. Third-party articles, patient records, API keys and original full-text downloads are not distributed in this package. Source versions and locators link back to the original reports.

The AI-use declaration and Funding statement for the manuscript remain for the author. This status record describes the resource preparation and verification state; it does not complete either manuscript statement.

The evidence resources were made public on 9 November 2026 with author authorization. Manuscript revision continues, and arXiv upload and submission remain paused. Author verification is incomplete. No additional open-content license has been selected.
