Preflight Check logo

Preflight Check

Community
ShaishavMaisuria
preflight-check

Deterministic desk-reject preflight linter for LaTeX paper submissions. Use it when a researcher asks "is my paper ready to submit", or mentions preflight, desk reject, submission check, page limit, anonymization, double-blind, missing checklist, or format/template compliance for a conference (NeurIPS, ICML, CHI, SIGSPATIAL, SIGMOD, ICDE, CVPR, LNCS...). Lints the .tex source against a machine-readable venue profile for documentclass and style-file options, margin/template tampering (geometry, savetrees, spacing hacks), anonymization leaks (author block, acknowledgments, repo links, first-person self-citations, \thanks, pdfauthor), required sections (NeurIPS checklist, ICML impact statement, AI-use acknowledgement, COI, CCS concepts/keywords), abstract length, keywords format, and page-limit risk. Every finding is reported with file:line. Runs bundled stdlib-only Python scripts; advisory only, never submits anything.

Overview

PublisherShaishavMaisuria
Repositoryresearch-paper-lifecycle-skills
Skill namepreflight-check
Stars
59
Forks
22
Bundled files
9
LicenseApache-2.0
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 9 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by ShaishavMaisuria on GitHub. Read the source before you install it.

Installation

Install the Preflight Check AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/ShaishavMaisuria/research-paper-lifecycle-skills/tree/main/skills/preflight-check
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/ShaishavMaisuria/research-paper-lifecycle-skills.git /tmp/research-paper-lifecycle-skills
mkdir -p .claude/skills
cp -r /tmp/research-paper-lifecycle-skills/skills/preflight-check .claude/skills/preflight-check
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Preflight Check in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Preflight Check on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Preflight Check is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Preflight Check

Run a deterministic desk-reject lint over a LaTeX submission before the authors hit "submit". Venues explicitly desk-reject — without review — for page-limit violations, template tampering, anonymization leaks, and missing mandatory sections. Every one of those is machine-checkable from the source. This skill catches them while there is still time to fix them.

When to use

  • "Is my paper ready to submit to NeurIPS / SIGSPATIAL / CHI / ...?"
  • "Check my paper for desk-reject risks" / "run a preflight"
  • "Did I anonymize this properly?" / "double-blind check"
  • "Am I over the page limit?" / "is my template compliant?"
  • Right after tailor-to-venue, and always before any actual submission.

Inputs

  1. The main .tex file of the submission (the one with \documentclass). \input/\include files are followed automatically.
  2. A venue profile: venues/conferences/<venue>-<year>.yml (schema in venues/schema.yml). If the venue has no profile yet, create one first with parse-cfp (or copy the nearest family file and fill it in).
  3. Optional but strongly recommended: a compiled .log or .pdf next to the .tex (same basename) so the page-count risk check can run.

Process

  1. Resolve the venue profile. Find the matching file under venues/conferences/. Confirm the year matches the target deadline cycle. Ask the user which track they are submitting to — page limits differ per track (e.g. SIGSPATIAL Research = 10 pages excl. references, Demo = 4 incl.). Sanity-check the merged profile with: python3 scripts/venue_profile.py venues/conferences/<venue>.yml --track <name>

  2. Re-verify critical facts against the live CFP — mandatory. Profiles are a starting point, never ground truth; a stale page limit causes the exact desk reject this skill exists to prevent. Fetch the profile's cfp_url and confirm: page limit + what it excludes, blind level, template invocation, required sections, and the deadline. If anything differs, update the profile YAML first and say so in the report.

  3. Run the full preflight:

    python3 scripts/run_preflight.py paper.tex \
        --venue venues/conferences/<venue>-<year>.yml --track "<track>"

    Or run checkers individually when the user asks about one dimension:

    ScriptChecksRun
    scripts/check_template.pydocumentclass + options vs profile, year-versioned .sty (NeurIPS/ICML/ICLR), geometry/savetrees/fullpage, \setlength on layout dims, \linespread<1, negative \vspace patterns, column count, page-limit risk from .log/.pdfpython3 scripts/check_template.py paper.tex --venue <yml> --track <t>
    scripts/check_anonymization.py\author/\affiliation/\email/\orcid/\thanks content, acknowledgments sections, funding/grant lines, identifying links (GitHub, personal pages) vs anonymized mirrors, first-person self-citations, hyperref pdfauthorpython3 scripts/check_anonymization.py paper.tex --venue <yml>
    scripts/check_sections.pyevery format.required_sections token: NeurIPS checklist, ICML impact statement, AI-use acknowledgement (ICDE), COI declaration, CCS concepts + \keywords, plus title/abstract/bibliography presencepython3 scripts/check_sections.py paper.tex --venue <yml>
    scripts/check_abstract.pyabstract word count vs venue bounds (or the 150–250 norm), single-paragraph rule, citations/URLs in abstract, keywords format (ACM/IEEE Index Terms/LNCS 3–6)python3 scripts/check_abstract.py paper.tex --venue <yml>

    Useful flags on every script: --json (machine-readable), --strict (warnings also fail), --track <substring>, --venues-dir <dir> (when the profile lives outside this repo), --no-inputs. Anonymization adds --force to scan even at single-blind venues.

  4. Interpret severities for the user.

    • ERROR — documented desk-reject grounds at this venue. Must fix.
    • WARN — judgment call (e.g. one negative \vspace is normal; eight is space compression). Walk through each with the user.
    • INFO — confirmations and pointers to manual checks. Do not pad the report with these; summarize. Exit codes: 0 = no errors, 1 = errors found (or warnings with --strict), 2 = bad arguments/missing files.
  5. Do the manual checks the linter cannot do. Source-level linting cannot see compiled-PDF metadata, where the references actually start, supplementary files, or submission-form fields (COI declarations at ICDE live in CMT, not the PDF). Work through references/manual-checks.md with the user.

  6. Fix and re-run until clean. Quote each finding's file:line, propose the minimal edit, apply it if asked, and re-run the affected checker. For the reasoning behind each check (which venue desk-rejects for what, with sources), see references/desk-reject-triggers.md.

Output

A findings report (text or --json) per checker or combined via run_preflight.py: severity, check id, file:line, message, summary counts, and a verdict (FAIL if any ERROR, PASS with warnings if only WARNs, else PASS). Present it to the user as a fix-list ordered by severity, then offer to apply fixes one at a time.

Worked example

A NeurIPS submission that still carries author identity and last year's style file. Running the combined checker:

python3 scripts/run_preflight.py paper.tex \
    --venue venues/conferences/neurips-2026.yml
=============================================================================
PREFLIGHT REPORT  paper.tex
venue: neurips-2026   track: main
=============================================================================

-- template --------------------------------------------------------------
   [ERROR] template/style-file-year             paper.tex:1 — uses neurips_2025.sty; NeurIPS 2026 requires neurips_2026.sty
   [WARN ] template/negative-vspace             paper.tex:312 — negative \vspace before figure; isolated, likely fine

-- anonymization ---------------------------------------------------------
   [ERROR] anonymization/pdf-metadata           paper.tex:14 — \hypersetup{pdfauthor=...} leaks the author into PDF metadata
   [WARN ] anonymization/author-block           paper.tex:22 — \author populated; the .sty hides it at submission, but scrub the source
   [WARN ] anonymization/self-citation          paper.tex:88 — "our previous work [12]" — rewrite in third person

-- sections --------------------------------------------------------------
   [ERROR] sections/missing:neurips-checklist   paper.tex — no NeurIPS checklist found; a missing checklist is a desk reject

-- abstract --------------------------------------------------------------
   clean.

=============================================================================
VERDICT: FAIL — fix ERRORs before submitting   (3 error(s), 3 warning(s), 0 info)
Profiles can go stale: re-verify limits/policies at https://neurips.cc/Conferences/2026/CallForPapers
=============================================================================

How to present this: lead with the three ERRORs as a numbered fix-list (neurips_2026.sty, drop the pdfauthor leak, add the checklist), each with its file:line and the one-line edit; then walk the WARNs as judgment calls. Re-verify the page limit and checklist requirement at the cfp_url before the user relies on the verdict, then re-run the affected checker after each fix.

Adapt to your discipline

Profiles can represent conferences, journals, and workshops in any field. Add venue YAMLs with your community's rules — the checkers only read the profile, so new disciplines need data, not code.

Guardrails

  • A clean preflight is necessary, not sufficient — never tell the user the paper "will not be desk-rejected"; say "no machine-checkable desk-reject triggers found".
  • Always re-verify page limits/deadlines/policies against the live cfp_url before the user relies on them (step 2 is not optional).
  • Never submit to any system on the user's behalf; stop at the report.
  • Never paste large portions of the user's paper into the report — quote at most the flagged line.
  • Citation problems are out of scope here: route them through verify-citations.

Memory

This skill uses the shared .paper-memory/ convention in the user's paper directory, following paper-memory-convention.md.

  • At start: read .paper-memory/lessons.md (and profile.yml for venue tier / constraints). Skip re-flagging issues already recorded and acted on; lead with any recurring desk-reject habits this author has (e.g. "you tend to compress layout near the page limit").
  • At end: append each finding worth remembering as one dated entry in the shared format - [YYYY-MM-DD] (preflight-check | <scope>) issue -> recommendation (use reflect-and-improve's reflect_log.py append, which dedupes and dates). Tag a repeat-across-papers habit recurring, a one-off this-paper.
  • Create .paper-memory/ on demand if absent and offer to add it to the project .gitignore. It is local-only; never upload it or copy it into this repo.

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Preflight Check AI skill do?

Deterministic desk-reject preflight linter for LaTeX paper submissions. Use it when a researcher asks "is my paper ready to submit", or mentions preflight, desk reject, submission check, page limit, anonymization, double-blind, missing checklist, or format/template compliance for a conference (NeurIPS, ICML, CHI, SIGSPATIAL, SIGMOD, ICDE, CVPR, LNCS...). Lints the .tex source against a machine-readable venue profile for documentclass and style-file options, margin/template tampering (geometry, savetrees, spacing hacks), anonymization leaks (author block, acknowledgments, repo links, first-p...

Why use Preflight Check on TypingMind?

Because you install it once and use it with any model. Preflight Check is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Preflight Check in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/ShaishavMaisuria/research-paper-lifecycle-skills/tree/main/skills/preflight-check. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Preflight Check?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Preflight Check?

As many as you like. As long as a model supports skills, you can use Preflight Check with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Preflight Check AI skill free?

Yes. It is published on GitHub by ShaishavMaisuria under the Apache-2.0 license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇