auto-paper-improvement-loop

Automating academic paper improvement involves running a two-round iterative loop using AI to review and fix LaTeX documents. This Skill automatically iterates on a compiled paper, performing structured reviews and recompiling updated PDFs to boost clarity and rigor.

To run an automated LaTeX paper polish loop, you need a compiled paper directory containing paper/main.pdf and paper/sections/*.tex files. The Skill uses these inputs to execute two rounds of reviewer-guided improvements and output updated PDFs.

Yes, AI can automatically iterate on a compiled LaTeX paper to fix issues. The workflow performs two rounds of reviewer-guided improvements using GPT-5.4 xhigh, generating updated PDFs like main_round1.pdf and main_round2.pdf, with an optional human checkpoint between rounds.

Crash-safe progress recovery for automated paper writing works by persisting state to a PAPER_IMPROVEMENT_STATE.json file. This allows the AI review and fix loop to resume gracefully without losing iterative improvements if the process is interrupted.

The best way to track AI paper improvements in LaTeX is by generating a detailed improvement log. This Skill automatically outputs a PAPER_IMPROVEMENT_LOG.md file documenting every structured review and fix applied across the two rounds of iteration.

A limitation of using AI for automated LaTeX paper improvement is its dependency on a pre-existing compiled paper structure with specific section files. It operates strictly as a post-writing polish loop and cannot generate the initial academic content from scratch.