mcpField
P
paper-section-writer

paper-section-writer

github.com/zhnnky329Documents1.1k38 installsUpdated 5w ago
Repository

About paper-section-writer

A skill for drafting mathematical modeling contest paper sections from finalized methods, results, and recorded human decisions. It traces numerical claims to frozen sources and keeps physical interpretation and contribution framing under human control.

Key features

  1. 01Prerequisite and source checks
  2. 02Claim mapping with provenance
  3. 03Drafting from finalized methods and results
  4. 04Human-owned interpretation and framing
  5. 05Consistency across figures and formulas
  6. 06Limitations and uncertainty included

Use cases

  1. 01Drafting a section from finalized contest results
  2. 02Tracing paper claims to frozen values
  3. 03Preparing text that matches verified figures and methods

Install

Copy the command and run it in a terminal.

$npx skills add https://github.com/zhnnky329/mathmodeling-skills --skill paper-section-writer

After installing, type/paper-section-writerin Claude Code to confirm it loaded.

Details

From the project README

Preconditions

  • rigor_profile is submission.
  • Final method explanation exists.
  • Final result analysis exists.
  • Solution package and current frozen numbers exist.
  • Required human claim-scope and physical/domain-meaning decisions are recorded.

If any prerequisite is missing, return to its producer rather than drafting around the gap.

Primary Sources

Use, in order:

  1. qx_solution_package_for_writer.md
  2. frozen_numbers.json
  3. qx_decisions.jsonl
  4. verified paper figures/tables
  5. final method explanation and robustness report for clarification

Do not hunt through raw experiment folders to invent a narrative.

Workflow

  1. Resolve the requested section and contest format.
  2. Build a claim map:
    • claim ID;
    • frozen value/source;
    • robustness support;
    • human decision ID;
    • figure/table reference;
    • limitation.
  3. Draft the method description to match the final explanation and code.
  4. Draft results with:
    • value and comparison;
    • human-confirmed physical/domain meaning;
    • uncertainty or robustness;
    • limitation and applicable scope.
  5. Mention the baseline and eliminated alternatives only when they explain a real decision.
  6. Use only Type 2–4 figures as appropriate; never place Type 1 diagnostics in the paper.
  7. Save paper/sections/qx.tex or the requested Markdown section.

Human-Owned Content

The AI must not originate:

  • why the method was chosen;
  • what the headline number means physically;
  • confidence and claim scope;
  • contribution framing.

Transcribe these from the decision ledger with provenance. If absent, invoke a compact choice card and stop the final draft until answered; do not fill the paper with repeated sentinels.

Rules

  • Every numerical claim must match frozen_numbers.json.
  • Do not overclaim against untested methods or populations.
  • Do not fabricate citations or causal meaning.
  • Avoid procedural diary prose and ceremonial detail.
  • Keep formulas, symbols, units, captions, and filenames consistent.
  • Do not create a new decision artifact.

Verification

  • Three writer prerequisites pass.
  • Claim map resolves all numbers and judgments.
  • Method, results, and figures match canonical artifacts.
  • Physical meaning and contribution are human-owned.
  • Limitations and uncertainty are visible.
  • No Type 1 figure appears.

Comments

Markdown supported