submit

submit — the pre-submission gate

Baca dalam bahasa Indonesia

v1.5.0 · download zip

Sweeps a manuscript for desk rejection risk — rejection by the editor before it ever reaches a reviewer, in roughly a ten-minute read, affecting 40–70% of submissions at reputable journals.

Most of the criteria have nothing to do with scientific quality. Good manuscripts are returned routinely because the scope missed or an ethics statement was left blank.

What it is actually for

1. The manuscript is ready and you want to know what will kill it

"I want to send this to [journal name]. Here are its author guidelines. Check it."

The skill runs in order, from cheapest and most lethal first. Journal scope before hyphens. The moment a fatal finding appears, it stops and reports — because polishing sentences in a manuscript aimed at the wrong journal is work that will be thrown away entirely.

What you must have ready: the target journal's author guidelines URL. Without it roughly half the risk cannot be assessed, and the skill will say so rather than guess — word limits and required statements differ between journals from the same publisher.

2. It came back without review and you do not know why

This is the situation that brings most people here, and the answer is often surprising: most desk-rejection criteria never touch scientific quality. Good manuscripts are returned because an ethics statement was blank, a word limit was exceeded, or anonymity leaked through file metadata.

3. You want to know whether the citations actually exist

Three statuses that other tools usually merge, and the difference matters:

Status	Meaning	What to do
VERIFIED	matches the index	—
MISMATCH	DOI is right, authors or year are off	the work is real; the bibliography entry is wrong
UNVERIFIED	not found in any index	a strong sign of a fabricated citation

And one check rarely found elsewhere: check_claim_fidelity.py asks whether the source actually says that — verbatim quotations absent from the source, concept attributions where not one of the words appears, cardinal claims ("reports three strategies") whose number is never stated.

If the source's full text is unavailable, that script stays silent. It never guesses.

When to use it

• The manuscript is considered finished and about to be sent
• It was rejected without review and you do not know why
• You are choosing or changing target journals
• You are assembling a submission package: cover letter, highlights, ethics statements, anonymisation

This is a gate, not a quality sweep

The difference from polish-manuscript and nulis is not just the checklist but the shape:

• Those two are improvement passes — linear, exhaustive, weighing all dimensions equally
• This one is a gate — sequential, cheapest and most lethal first. One fatal finding makes the rest irrelevant: stop, report, do not carry on sweeping.

No point polishing hyphens in a manuscript aimed at the wrong journal.

What you must prepare

1. The manuscript (.docx/.tex/.md)
2. Target journal + its author guidelines URL — non-negotiable; without it roughly half the risk cannot be assessed
3. Review model: double-blind or single-blind? Was it ever a preprint?
4. Names and affiliations of all authors

The author guidelines are actually read, not inferred from publisher habit — word limits and required statements differ between journals from the same publisher.

Five scripts

Script	Checks
sweep.py	consistency within the manuscript: word count, seven required statements, two-way citation cross-check, anonymity leaks, orphan figures, leftover TODOs/tracked changes
verify_refs.py	references against the outside world — CrossRef, OpenAlex, PubMed
check_claim_fidelity.py	does the source actually say that?
prisma_cascade_check.py	PRISMA cascade arithmetic
check_prisma_figure.py	manuscript versus PRISMA figure

---

← Back · Installation · Workflow · FAQ