- Home
- Skills
- Research & Analysis
- paper-audit
Works with the AI Agents you already use
Paper Audit
Audit and polish academic manuscripts — figures, equations, typography and layout — with rendered visual checks.
$1
paper-audit
Example session with this skill installed
Audit my manuscript's figures, equations and typography before journal submission.
- Read your context and instructions
- Compiled the paper-audit
What the buyer receives
A complete pre-submission audit package for a 22-page LaTeX manuscript headed to a journal — five deliverables that leave the science untouched and the presentation spotless:
1. Categorized audit report
audit-report.md — every finding sorted into four clearly separated buckets so the buyer knows exactly what kind of change each one is:
Layout / rendering: three figures rendered past the text margin in the compiled PDF; one two-column figure landing in the middle of a paragraph instead of at the top of the page; an oversized table breaking the page bottom with its last row orphaned onto the next page. Each finding names the page, the figure or table, and the fix.
Typography / punctuation: 14 instances of straight quotes that should be curly in body text; two en-dashes used where em-dashes belong; missing non-breaking spaces before all 31 citations (LaTeX renders ~\cite correctly, the draft used plain spaces); inconsistent serial commas across sections; three occurrences of "et al." typeset without the thin space the journal style requires.
Clarity: five figure captions that describe what the figure shows but not what it means — each rewritten with a one-sentence takeaway added, clearly marked as suggested text the buyer can accept or edit. Two paragraph transitions flagged where a signpost sentence would help the reviewer follow the argument.
Substantive flags (separate, no changes made): two places where the text claims a result the figure does not obviously show, and one equation reference that points to the wrong numbered equation. These are flagged for the buyer to verify — the audit never "fixes" science, it only reports what it found.
2. Before/after renders of fixed figures
figures/before-after/ — for each figure the audit repaired (cropped whitespace, re-exported at journal DPI, fixed font embedding, re-aligned subfigure labels), a side-by-side PNG pair: the figure as it appeared in the submitted PDF versus the corrected version. The buyer sees exactly what changed before touching their source, and every correction is described in plain language next to the images.
3. Equation numbering and cross-reference checks
equations.md — a full pass over the 47 numbered equations: every \label / \ref pair verified, two broken references repaired (they pointed to equations that had been renumbered in an earlier revision), one unnumbered display-math block that was referenced in text promoted to a proper numbered equation, and a check that all in-text references use the journal's required format ("Eq. (3)" vs. "equation 3" standardized throughout). The compiled PDF was re-rendered to confirm no "??" undefined-reference markers remain anywhere in the document.
4. Caption consistency pass
captions.md — all 9 figure captions and 6 table captions normalized to the journal's style: sentence-case vs. title-case made consistent, terminal periods added or removed uniformly, subfigure labeling aligned ((a), (b), (c) ordering verified against the actual panel layout), and every caption checked for a matching in-text citation — one figure was never referenced in the body at all and is flagged with a suggested insertion point.
5. Clean bill of typographic health with minimal source diffs
fixes.diff — the actual source changes, kept as small as possible: whitespace fixes, punctuation swaps, reference repairs, and caption edits, each in a separate commented hunk. The diff is the deliverable the buyer applies, and it touches nothing else — no reformatting of unrelated paragraphs, no package changes, no rewording of claims. Applying it is a git apply away, and a final compiled PDF proves the manuscript still builds cleanly with zero new warnings.
How it was made
The manuscript was compiled from source and inspected in its rendered form — the audit works from what the reviewer will actually see, not just the .tex file. Figures were re-rendered at publication DPI, the full document was rebuilt after fixes, and every change was verified against the rendered PDF rather than assumed from source edits.
What makes this different
Most "proofreading" is spellcheck plus vibes. This is a repeatable, evidence-backed inspection: rendered visual checks, categorized findings, substantive issues kept strictly separate from presentation fixes, and the smallest possible diff. The buyer gets a manuscript that looks like it was prepared by someone who has submitted a hundred papers — because the checks encode exactly that experience.
Connects securely to your tools. The creator never sees your data.
About this skill
A repeatable pre-submission audit for scholarly manuscripts: figures, equations, captions, typography, punctuation and page layout, backed by rendered visual checks — not just source inspection. Content-preserving by default: claims, data, citations and figure meaning stay intact; presentation changes only on request or for unambiguous typographic errors. Works with LaTeX, Word, PDF and Markdown projects.
How to use
Skills on Agensi run inside your AI agent. Copy the prompt, paste it into your agent, and it connects to Agensi and installs this skill for you. If a payment is needed, your agent shows it to you first.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill passes our trust checks below.
Security scanned
Passed our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scanned
Verified clean today
- Passed all security checks, Safe to install