Skip to content
Skills Library
Science & research · Review & audit

check-draft

/check-draft

Run chosen specialists over a draft of your own - statistics, the paper's own numbers and methods reporting by default - with no reviewer track and no editor. Use when asked to check, audit or sanity-check a draft or manuscript before submission, rather than to peer review it.

Licence
MIT
Last repository change
13.11.2026

Assessment and security

96/ 100Class: good
Format25 / 25
Frontmatter with name and description, valid name, description up to 1,024 characters, name matching the folder.
Discoverability25 / 25
Description explains when to use the skill, has a reasonable length and uses third person.
Content31 / 35
At least 150 words and at most 500 lines, headings, step lists, examples or code.
Hygiene15 / 15
No placeholders such as TODO, no Windows paths, with a licence.
Riskno signals detected in the scan
Files in this package
1
Scripts
0
HTML comments
0
URLs in SKILL.md
0
Deprecation

No signals detected. Repository changed on 13.11.2026, without references to retired models or APIs.

Licence

MIT, open: use and modify while retaining author attribution.

What it does and when to use it

Run chosen specialists over a draft of your own - statistics, the paper's own numbers and methods reporting by default - with no reviewer track and no editor. Use when asked to check, audit or sanity-check a draft or manuscript before submission, rather than to peer review it.

Original author description

Run chosen specialists over a draft of your own - statistics, the paper's own numbers and methods reporting by default - with no reviewer track and no editor. Use when asked to check, audit or sanity-check a draft or manuscript before submission, rather than to peer review it.

How it works

SKILL.md has no level-two headings · 51 lines, 430 words

SKILL.md preview
---
name: check-draft
description: Run chosen specialists over a draft of your own - statistics, the paper's own numbers and methods reporting by default - with no reviewer track and no editor. Use when asked to check, audit or sanity-check a draft or manuscript before submission, rather than to peer review it.
---

# Check a draft

This is the author's pass, not peer review. The specialists audit their own
dimension and nothing pools their findings into a verdict, because a draft of
your own does not need one — it needs the list.

1. Take the PDF path, the run folder and, if the request names them, the
   tracks. The default is `stats`, `numbers`, `methods-reporting`.
   The six available are `stats`, `figures-tables`, `methods-reporting`,
   `citations`, `numbers`, `field`. Each audits one dimension and nothing
   another one owns, so naming a track decides what goes unchecked: `numbers`
   is every printed quantity and the arithmetic between them, `stats` the
   inference drawn from them, `figures-tables` what a panel shows,
   `methods-reporting` what is absent from the methods, `citations` the
   references, `field` what this kind of study is required to do.
2. Call the `Workflow` tool with `name` set to `pra:review` (if that name is
   not found, try `review`) and `args` set to:

   ```json
   {"pdf": "<the PDF path>", "out": "<the run folder>", "arm": "draft-check",
    "tracks": ["stats", "numbers", "methods-reporting"]}
   ```

   Add `"specialist_model": "<a model>"` only when the request names one; left
   out, every specialist runs on Sonnet. If the workflow refuses to start
   because a model is not available on this login, report that line as it came
   and do not retry with another model.

3. Wait for it, then report what it returns, and point at the per-track audits
   the run wrote: `<run folder>/review/audit_<track>.md`, one per track. An
   audit may carry a **Seen, not mine** line: something that specialist noticed
   in another lane and did not file. Pass those on as they are written — with
   the track that owns them unrun, nothing else looked.

With `citations` among the tracks, the run also keeps a citation ledger —
`<run folder>/review/citations/ledger.jsonl`, one row per reference and one per
in-text use, with each reference looked up and each use judged against the
source before the specialist reads the table. It is the slowest track and the
only one that needs the network. Report its counts as the run gives them, and
keep "could not be checked" apart from "not found": a lookup that failed says
nothing about whether a reference exists.

The `draft-check` arm runs no core reviewer and no editor, so every finding
stands on its own and all of them are in the review's body. There is no
verdict, and you should not invent one — if the request wants a judgement on
whether the paper would be accepted, that is `peer-review-manuscript` with the
`full-two-track` arm.

Do not read the PDF yourself, write files, or add findings of your own.

Versions

The database has no other skills named check-draft.

Database snapshot: 13.11.2026 · package version 3bfadb90bd6b (SHA-256 digest of the entire package)

Downloads in the last 30 days: 0 · ranking

Report a problem with this skill: harmful content, incorrect descriptions or an author’s removal request.

Similar skills

  • 100paper-review
    Use this skill for \"review this paper\", \"review this manuscript\", \"peer review\", \"review my paper\", \"critique this manuscript\", \"review this submission\", \"give me feedback on my paper\", \"check my methods\", \"review my statistics\", \"review as a peer reviewer\",…
  • 100paper-review
    Use this skill for \"review this paper\", \"review this manuscript\", \"peer review\", \"review my paper\", \"critique this manuscript\", \"review this submission\", \"give me feedback on my paper\", \"check my methods\", \"review my statistics\", \"review as a peer reviewer\",…
  • 100peer-review-manuscript
    Run a full two-track peer review of one manuscript PDF - N core reviewers and six specialists, then an editor who pools their findings into one review. Use when asked to peer review, referee or critique a manuscript, or to run PeerReviewAgents on a PDF.
  • 100peer-review
    Reviews a manuscript, proposal, analysis or result the way a careful referee does, reading the whole artifact, checking the methods against the claims, the statistics against the design, the figures against the numbers, and reporting BLOCKING issues separately from observations,…
  • 100reviewer-2-simulator
    Critiques your paper draft as a skeptical reviewer would. Use when asked to review a paper draft, find weaknesses in a paper, prepare for peer review, anticipate reviewer criticism, or stress-test research before submission. Identifies weak claims, missing baselines, unclear…
  • 100paper-review
    Multi-agent academic paper review system. Use this skill whenever the user asks to review, critique, improve, or get feedback on an academic paper, manuscript, preprint, or draft. Triggers include: 'review my paper', 'critique this manuscript', 'find weaknesses in my paper',…

Feedback and reports

Your message goes to Bartosz Sokoliński. An email address is needed only for a reply. See the privacy policy.