katz

Read the paper and decide which issue spotters to use, removing irrelevant ones and optionally adding custom ones

Run EDSL-parallelized issue finding across all sections of a katz-registered manuscript

Bootstrap a full paper review using katz

Send paper figures to vision-capable models for feedback on clarity, design, and presentation

Review open katz issues, investigate them against the manuscript and code, and record findings

Read the paper and file issues for problems found in the manuscript

Review the paper's repository — code, data, and analysis scripts — against manuscript claims

Collaboratively create GitHub issues from review findings

Generate actionable suggestions for confirmed issues and low-grade evaluations

Evaluate the paper against enabled criteria, writing candid narrative responses

Synthesize investigated katz issues into a narrative referee report

Convert the paper source to canonical markdown and register it with katz

Chunk the registered paper into logical sections and add them to the katz paper map