AISTATS Skills

Twelve agent skills for AISTATS papers at the AI, statistics, and machine-learning interface. The pack focuses on OpenReview submission readiness, statistical evidence, double-blind supplementary material, reproducibility checklists, text-only author-reviewer discussion, PMLR camera-ready publication, and in-person presentation planning.

Official basis checked on 2026-06-01: AISTATS 2026 Call for Papers, OpenReview conference group, PMLR proceedings pages, AISTATS Code of Conduct, and AISTATS / PMLR author publication conventions. See resources/official-source-map.md for the exact source map.

Skills

• aistats-submission - audit OpenReview, page limits, anonymity, supplement, checklist,

dual submission, and reviewer-volunteer requirements.

• aistats-author-response - draft concise, anonymous, text-only author-reviewer discussion

replies without forbidden links or unsupported new claims.

• aistats-camera-ready - prepare PMLR camera-ready files, de-anonymization, metadata,

registration, in-person presentation, and artifact release.

• aistats-artifact-evaluation - package code/data/proofs/logs as anonymous supplementary

evidence or public post-acceptance artifacts.

• aistats-reproducibility - strengthen checklist, assumptions, proofs, datasets,

hyperparameters, seeds, compute, and uncertainty evidence.

• aistats-supplementary - organize appendices, proof details, code/data archives,

simulation scripts, and anonymized artifacts.

• aistats-review-process - reason about OpenReview reviews, discussion, meta-review,

confidentiality, and PMLR outcomes.

• aistats-writing-style - revise for compact AI-statistics framing, assumption clarity,

theorem/experiment balance, and 8-page compression.

• aistats-related-work - position against AI, ML, statistics, arXiv, workshop, concurrent,

and archival literature.

• aistats-experiments - audit simulations, baselines, statistical tests, uncertainty,

ablations, seeds, hyperparameters, compute, and data handling.

• aistats-workflow - manage the calendar from fit through abstract, full paper, supplement,

reviews, discussion, decision, camera-ready, and presentation.

• aistats-topic-selection - decide whether a project is AISTATS-shaped or better routed to

NeurIPS, ICML, ICLR, UAI, COLT, JMLR, a statistics journal, or an application venue.

Maintenance notes

• Reopen current-cycle official sources before deadline-sensitive advice.
• Treat AISTATS 2026 facts as historical anchors, not permanent rules.
• Page limits, supplementary deadlines, discussion rules, PMLR forms, registration, and

presentation obligations can change by cycle.