pre-review

name	pre-review
description	Run an evidence-first pre-submission review across design, data, methods, statistics, results, claims, reproducibility, ethics, and journal-pack rules. Use after manuscript preflight to create a prioritized revision board without predicting editorial acceptance.

Pre-Review

Run a rigorous author-side review before packaging. This Skill does not impersonate an external reviewer and does not predict acceptance.

Required inputs

Require a readable manuscript or stable excerpt and a valid evidence-ledger.json. Use supplied tables, figures, supplements, protocol, analysis outputs, and data audit when available. If a required artifact is missing, invalid, or blocked, set status to BLOCKED or PARTIAL_REVIEW; never infer a pass.

The extension may inject one active journal pack. Use it only for explicit rule checks. Without a valid active pack, mark journal-specific fit and package rules not assessed.

Evidence rules

Use missing evidence, negative evidence, and uncertain precisely. Anchor every finding. Never invent analyses, citations, journal rules, author intent, or source access. Never infer misconduct or AI authorship from writing style.

Review dimensions

Review each dimension independently:

1. question, contribution, and claim boundary;
2. design, population, comparator, and bias;
3. data provenance, measurement, and missingness;
4. statistical analysis, uncertainty, multiplicity, and sensitivity;
5. results, tables, figures, and numerical consistency;
6. claim-to-evidence alignment and interpretation;
7. reproducibility, reporting, and availability;
8. ethics, registration, declarations, and permissions;
9. active journal-pack rules, when available.

Mark coverage reviewed, partial, or not assessable. For readable numerical claims, create a numeric consistency ledger across the abstract, body, tables, figures, supplements, and supplied analysis outputs. Record conflicts and unavailable comparisons; do not infer values from an unreadable artifact. Journal fit means only the rules in the active pack. Without a pack, it is not assessed, not a general journal recommendation.

Findings and decision

Use stable finding IDs and priorities blocker, major, or minor. Keep evidence, issue, impact, action, and uncertainty distinct.

Choose:

• READY_FOR_PACKAGE: no blocker or major finding remains and required areas were assessable;
• REVISION_REQUIRED: review is sufficiently complete and substantive revisions remain;
• BLOCKED: a required input, traceability contract, or central claim cannot be evaluated responsibly;
• PARTIAL_REVIEW: useful findings exist but material dimensions were not assessable.

No unresolved item may be silently converted into READY_FOR_PACKAGE.

Call record_checkpoint with the strongest blocker, uncertainty, or stop rule. Prefix the statement with [pre-review].

Output

Write pre-review-report.md from ../../templates/pre-review-report.md, including coverage, prioritized findings, numeric consistency, revision order, status, and human judgments still required.