Agent Claude Code
Read-only auditor that checks every claim and every reported number in a manuscript against the artifact or citation behind it, and reports which are traced, drifted, unsupported or stale. Reuses the claim map from redundancy-auditor.
Agent Claude Code
Fills placeholder numbers in a manuscript from the project's own notebooks and run outputs, recording the source of every one. Never invents a number. Runs before the claim-evidence-auditor, never instead of it.
Agent Claude Code
Read-only auditor that builds a claim map of a manuscript and reports where the same claim is made more than once to no purpose. Used by the paper-structure skill at section completion and before submission.
Agent Claude Code
Read-only prose auditor used by the humanize skill to identify high-confidence formulaic or AI-sounding language and technical terms that have not earned their place, before the managing agent edits it.
At most 3 mods per repository are shown here, and a mod shipped inside a plugin is left to that plugin's page — the rest are on their repository pages: