- Home
- Skills
- Research & Analysis
- Paper Kit
Works with the AI Agents you already use
Paper Kit
Mock-review your paper with a journal-specific reviewer brief and fix scope, structure and numbers the way real reviewers flag them — rules distilled from actual rejections and revisions.
$9
Paper Kit
Example session with this skill installed
Review my paper for IEEE Transactions on Instrumentation and Measurement before I submit. The manuscript is in ~/paper.
- Read your context and instructions
- Compiled the paper kit
Before reviewing anything myself, I wrote a reviewer brief for this journal (paper-kit review.md section 1) and will hand it to a fresh subagent with no other context, so the review is independent of whoever wrote the text.
Reviewer brief
You are a reviewer for IEEE Transactions on Instrumentation and Measurement. The editor assigned you
this manuscript, and what you deliver is an accept/reject recommendation with its reasons. You are
not here to help the authors. Do not trust the authors' summary; read the manuscript text and the code
that produced its numbers yourself.
Review criteria, in this order:
- Is this a contribution to instrumentation and measurement, or a method paper in another field?
- Does the validation overlap with what is validated (circular reasoning)?
- Is any claim wider than the data shown?
- Is the measurand defined unambiguously?
- Are influence quantities identified and separated?
- Are the measurement floor and resolution reported separately from the measured values?
- Is what is called a bound really a bound, and do its assumptions hold on the data used?
- Is it reproducible? Is it useful to readers in this field?
Decision ladder: accept / minor revision / major revision / reject. No scores.
Read these (absolute paths):
- /home/user/paper/main.tex — manuscript
- /home/user/paper/numbers.tex — number macros used in the text
- /home/user/paper/results/ — result tables
- /home/user/paper/scripts/make_tables.py — produces the tables and numbers.tex
Tracing a number: a value in the text is a macro like \rmsErr; find it in numbers.tex, then find
the line in make_tables.py that writes it, then the row in results/ it was computed from. Recompute
at least the numbers behind the main claims.
Not a target for findings: typesetting, typos, reference formatting, and the numbers in Table 4,
which are being recomputed now.
Constraints the authors cannot change: one sensor model only; no access to a reference-grade
instrument. Judge whether the claims stand under these constraints, not whether more hardware would
help.
Output: for each major finding (a) where — sentence or file and line, (b) why it is a problem,
(c) what would resolve it. Then minor findings, a one-line recommendation, and a one-line scope
verdict.
Do not invent defects. If there are no major findings, say so. If you judge the paper out of scope
for this journal, say so — that is the most important finding.
Next steps
- Send this brief to a fresh reviewer and wait for its recommendation.
- Pair each major finding with its paragraph (style.md section 0) and mark it fix / do not fix / needs measurement.
- Fix only those paragraphs, then run the next round with a new reviewer. Stop when the reviewer says no revision is needed.
Connects securely to your tools. The creator never sees your data.
About this skill
Papers rarely get rejected for grammar. They get rejected because the claim is wider than the data, the validation is circular, the contribution is unclear, or the paper does not fit the journal. Paper Kit makes your agent check for exactly those things before a reviewer does.
What it does
- Mock review in rounds: the agent writes a reviewer brief for your target journal (its review criteria, its decision ladder, what to read, how to trace each number back to the script that made it) and hands it to a fresh reviewer with no prior context. You get a recommendation and reasons, never a score.
- Fix only what was raised: each finding is paired with its paragraph and judged fix / don't fix / needs measurement. No whole-manuscript rewrites.
- Real reviews and decisions: turns reviewer reports, editor decisions and co-author PDF comments into a per-finding fix plan.
- Style and scope: claim scope, defensive hedging, "reads like AI" phrasing, introduction and contribution structure, figures.
- Pre-submission machine checks: numbers match the generating scripts, no Type 3 fonts, no undefined references, page limits.
- Any journal: give a journal name and the agent builds that journal's standard from its author guidelines and sample papers.
What's inside
- SKILL.md: router that picks the right part for the request
- review.md: round structure and the reviewer brief, how to read real reviews, when to stop
- style.md: scope, structure, sentences, figures, machine checks, building a journal standard
- examples/reviewer-brief.md: a filled-in brief for an instrumentation and measurement journal
The rules come from real rejections and revisions on journal submissions, not from a style guide.
Works best with
An agent that can read your LaTeX source, result tables and scripts, and can start a fresh subagent (or use another model) as the independent reviewer. Text only: no scripts to install, no network calls.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Research & Analysis.
Does not perform final submission. Cannot verify author affiliations or funding facts. Requires user-provided data for specific journal standards.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by 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 scan
Security scan passed 2 days ago
- Software
- PopplerLaTeX