agentleFS
LocloveLoclove/cumcm-skills/skills/cumcm-reviewer/SKILL.md

cumcm-reviewer

Skill4 starsChanged 3 months agoMITView on GitHub

Simulate CUMCM judge review and red-team audit for mathematical-modeling papers, models, code, data loops, figures, citations, formulas, originality, the enhanced 25--30-page Song-font profile, Nature-style argument flow, figure-text integration, and final submission readiness. Use when the user asks to审稿,自审,模拟评委,查漏洞,终稿风险, or review a contest paper before submission.

The file

We split this skill into the rules it gives an agent

6 rules. About 0 are common, 5 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈5

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Use scripts/review_matrix.py for a deterministic checklist when a draft file is available.

    line 16
  2. Distinctive:

    Do not invent evidence or line numbers.

    line 48
  3. Distinctive:

    Do not downgrade broken reproducibility, missing citations, identity leaks, or invalid formulas to minor comments.

    line 49
  4. Distinctive:

    Do not accept filler, enlarged layout, repeated figures, or background padding as a valid way to reach 25 pages.

    line 50
  5. Distinctive:

    Do not accept visually polished figures that lack a body anchor, nearby interpretation, consistent terminology, or source-data linkage.

    line 51
  6. Distinctive:

    Do not rewrite the paper while reviewing unless the user asks for a repair pass.

    line 52

≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

CUMCM Reviewer

This skill reviews from the perspective of a contest judge. It should be sharp, concrete, and evidence-grounded.

Workflow

  1. Load ../cumcm-shared/references/quality-gates.md, ../cumcm-shared/references/format-spec-2026.md, and ../cumcm-shared/references/enhanced-writing-profile.md.
  2. Build an immutable review packet: problem requirements, paper draft, data/code inventory, figures, references, and format constraints.
  3. Produce three independent reviewer lenses: mathematical modeling, data/code reproducibility, and paper/format presentation. The third lens must audit 25--30 body pages, Song typography, paragraph jobs, claim-evidence proximity, and figure integration.
  4. Freeze the three reports, then synthesize blocking issues.
  5. Use scripts/review_matrix.py for a deterministic checklist when a draft file is available.

Output Contract

Review setup
- Input scope:
- Missing materials:

Reviewer 1: Modeling validity
- Major concerns:
- Minor comments:
- Blocking issues:

Reviewer 2: Data/code/results loop
- Major concerns:
- Minor comments:
- Blocking issues:

Reviewer 3: Paper, figures, citations, format
- Major concerns:
- Minor comments:
- Blocking issues:

Post-review synthesis
- Must fix before submission:
- Should fix if time allows:
- Strengths to preserve:

Red Lines

  • Do not invent evidence or line numbers.
  • Do not downgrade broken reproducibility, missing citations, identity leaks, or invalid formulas to minor comments.
  • Do not accept filler, enlarged layout, repeated figures, or background padding as a valid way to reach 25 pages.
  • Do not accept visually polished figures that lack a body anchor, nearby interpretation, consistent terminology, or source-data linkage.
  • Do not rewrite the paper while reviewing unless the user asks for a repair pass.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.