The Lab is not recruiting decorative contributors. These are bounded external verification assignments tied to specific research objects. A completed assignment creates an attributable review record and, where the evidence standard is met, changes the validation state of that release.
Independently inspect one official release from each of five institutions, compare source capture and transformations, include the specified conflict-affected Palestinian pair, and document discrepancies before any repair.
Completed audit template, source links checked, discrepancies, environment or process notes, attribution consent.
Code the frozen validation sample independently without seeing founder labels, then compare results only after submission.
Completed coding file, agreement statistics, disagreement log, attribution consent.
Audit a bounded sample of official-source ingestion, normalization and classification decisions and report any disagreements or coverage gaps.
Sample audit sheet, source comparison, disagreement log, limitations note.
After live delivery, use only the public package in a clean environment to recover the published teaching outputs and log any friction or divergence.
Reproduction log, environment, recovered outputs, discrepancies.
A real person completes the work and can be attributed in the review record unless a documented safety reason requires limited public identification.
The reviewer is not the primary producer of the object being validated and relevant conflicts are disclosed.
The work leaves an artifact: audit sheet, coding file, reproduction log, discrepancy record or methods memorandum.
The result maps to a defined validation field. Informal praise and private comments do not change release status.
Accepted work records the contributor, contribution type, affected artifact, date and evidence location. Authorship follows documented intellectual contribution; review alone is not authorship. Advisory appointment is also separate and is listed only after explicit acceptance.
Reviewers can use the bounded task specification, reviewer packet, validation dashboard and machine-readable assignment register. Open a repository issue before substantial work so scope and evidence requirements are frozen before review begins.
Methods & Reproducibility, Regional Data, and Open Research & Governance remain open advisory roles. They are not used as substitute evidence for release-level review.