agentsclimarketplace

Facct submission

Skill brycewang-stanford/Awesome-Journal-Skills/FAccT-Skills/skills/facct-submission

Use when auditing an ACM FAccT submission for OpenReview readiness — the mandatory abstract-registration gate with focus-area selection, the acmart anonymous/review PDF, the 14-page (+1 endmatter) budget, mutual anonymity, the required Generative AI Usage Statement and optional ethics/adverse-impacts statements, archival vs non-archival choice, and desk-reject triage before the deadline.From its SKILL.md

Install
npx -y skills add brycewang-stanford/Awesome-Journal-Skills --skill facct-submission

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

SKILL.md

6.7 KB, ~1.5k tokens by cl100k_base, as published. Nobody here has run it

FAccT Submission

Run this audit before uploading to the FAccT OpenReview site. FAccT papers are archival ACM proceedings in an interdisciplinary responsible-AI venue, reviewed mutually anonymously by a mixed CS+law+social-science pool. Every number below was read from the FAccT 2026 CFP and Author Guide on 2026-07-09 via search renderings (see resources/official-source-map.md); treat them as a one-cycle snapshot and reopen the live call first.

The two-step deadline (abstract gate first)

FAccT separates abstract registration from paper submission, and the gate is hard:

  • Abstract registration locks title, abstract, authors, and focus areas roughly a week early (FAccT 2026: 8 January 2026). Authors who do not register by this date cannot submit a paper — the focus areas drive reviewer/AC matching in the mixed pool.
  • Paper submission uploads the anonymized acmart PDF (FAccT 2026: 13 January 2026).

Register with the real title, abstract, and honestly chosen focus areas, not placeholders — a mismatch between your registered focus areas and your actual contribution mis-matches you to reviewers from the wrong discipline, the most damaging thing that can happen at an interdisciplinary venue.

Format and page budget

  • ACM acmart template, built anonymized for review, e.g. \documentclass[manuscript,screen,review,anonymous]{acmart} (single-column review format). The camera-ready later uses the sigconf proceedings format.
  • Page budget (verify per cycle): up to 14 pages excluding references (most papers 12-14), plus one additional content page for endmatter statements; references unlimited. Revised and accepted papers may use 15 content pages.
  • Do not modify margins, font, or spacing to gain space — editorial compression, not template tampering, is how you recover room.

Mutual-anonymity sweep

FAccT review is mutually anonymous: authors and reviewers are hidden from each other. Because FAccT papers often name systems, communities, and field sites, the leak surface is wide:

# Mechanical pass on the submission PDF and any anonymized artifact archive
pdfinfo paper.pdf | grep -Ei 'author|creator|producer'
pdftotext paper.pdf - | grep -nEi 'github\.com/[a-z0-9-]+|osf\.io|acknowledg|funded|grant|IRB #' | head
grep -rniE 'our (lab|group|team|university)|@[a-z0-9.]+\.edu' artifact/ --include='*.md' | head
unzip -l artifact.zip | grep -Ei '\.git/|\.DS_Store|/home/|/Users/' | head

The FAccT-specific leaks: a positionality statement left in the anonymous PDF (it is identity-revealing by design — remove it), acknowledgements or funding, an IRB protocol number, a named field site or partner organization that identifies the team, and a data-availability link to a personal repository.

Required and optional endmatter (do not skip)

  • Generative AI Usage Statement — REQUIRED. State whether and how generative AI was used to prepare the manuscript. FAccT prohibits using generative AI to wholesale-generate text. A missing statement is a compliance defect independent of the paper's merit.
  • Optional but expected where relevant: Ethical Considerations and Adverse Impacts statements — include them anonymized; they are read seriously.
  • Withhold until acceptance: Positionality, Author Contributions, Acknowledgements, Competing Interests — all identity-revealing; they must not appear in the anonymous submission.

Archival vs non-archival

Choose deliberately: archival (the default) publishes in the ACM DL proceedings; non-archival presents at the conference without proceedings publication (useful for work headed to a journal or already-published elsewhere). All accepted papers present in person regardless. Confirm the current non-archival mechanics on the Author Guide (待核实 per cycle).

Desk-risk triage

Finding at audit timeSeverityReal fix
Abstract/focus areas not registered by the earlier dateNo submission slot existsNothing fixes this post-deadline — calendar it now
Main text over the 14-page budgetDesk-reject-gradeCut or move to appendix; endmatter page does not absorb body text
Template altered (margins, shrunk font)Desk-reject riskRecompile clean, recover space editorially
Missing Generative AI Usage StatementCompliance defectAdd it before upload
Positionality/acks left in the anonymous PDFAnonymity violationRemove; restore only at camera-ready
Focus areas mismatched to the actual contributionWrong-discipline reviewersFix at registration; realign before the abstract gate
Identity leak (field site, IRB #, repo, funder)Anonymity violationRe-anonymize and re-host

Final-week order of operations

  1. Freeze the body early; endmatter statements can be drafted in parallel.
  2. Register title/abstract/authors/focus areas well before the abstract gate.
  3. Write the Generative AI Usage Statement; draft Ethical Considerations and Adverse Impacts.
  4. Run the mutual-anonymity checks on the final PDF and archive — and confirm positionality/acks are out.
  5. Fill every OpenReview field — focus areas, conflicts for every coauthor's institution and recent collaborators, archival/non-archival choice — a day early; late conflicts are the classic midnight failure.
  6. Re-download the uploaded PDF and read it cold to confirm it is the file you meant.

Reverify each cycle

  • The abstract-gate and paper deadlines and their time zone.
  • The page budget (14 vs 15, endmatter-page counting) and the exact acmart options.
  • Whether the Revise decision, OpenReview platform, and Generative AI Usage requirement persist (all introduced in 2026), plus non-archival mechanics and any APC/fee-waiver policy (待核实).

Output format

[FAccT submission status] ready / blocked / needs work
[Registration] title/abstract/authors/focus-areas locked by the earlier gate? yes/no
[Format] pages used (body/endmatter/refs), acmart anonymous-review compliance
[Anonymity] clean / leaks: <where> / positionality+acks withheld? yes/no
[Endmatter] Generative AI Usage present? ethics/adverse-impacts included?
[Archival choice] archival / non-archival, deliberate?
[Fix queue] <ordered, with owners and dates before the deadline>

What ships with it

Read from the repository

Just SKILL.md. No reference files, no scripts.

Keep looking

Skills are one crate of 325,949. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.