kill-argument

Stress-testing a LaTeX theory paper involves running an adversarial critique to generate a worst-case rejection argument, then adjudicating that attack point-by-point against your source text. This reveals whether your theorems and proofs truly support your claimed scope.

Adversarial review specifically decomposes a worst-case rejection into atomic points and labels each as answered_by_current_text, partially_answered, or still_unresolved. This classification exposes scope overclaims by validating whether your theorems and assumptions fulfill actual proof obligations.

Rebuttal preparation requires identifying the most lethal critique and matching it against existing source text. By generating an attack memo and classifying each point by its resolution status, you get a machine-readable KILL_ARGUMENT.json for targeted downstream planning and fixes.

An independent reviewer thread reads your relevant .tex and .bib files to locate citation-evidence gaps. It runs two fresh adversarial passes against your current text, isolating claims that lack supporting evidence and flagging them as still_unresolved for recommended fixes.

Adversarial critique is most effective for theory papers with substantive theorem, lemma, or proposition material. Papers without formal proof obligations may not benefit fully, since the adjudication process specifically targets headline scope dependencies tied to mathematical assumptions.

A point-by-point defense check requires your paper directory containing relevant .tex and .bib files, with an optional main.pdf. The process locates the LaTeX entry point, reads the source files, and outputs KILL_ARGUMENT.md plus KILL_ARGUMENT.json directly into that directory.