Skip to content
Back to skills

Academic Writing

ASecurity

Use when drafting or revising manuscript text that must project graph-backed scientific arguments into paper prose across Introduction, Related Work, Results, Discussion, Methods, figures, tables, source data, and citation handles.

  • 2 stars
  • 0 votes
  • 0 copies
  • 1 view
  • Added September 2, 2026
ai-agentsgonode

Security analysis

A100/100

Pro scans all 5 files and shows the line behind each finding

Scanned September 20, 2026

npx -y skills add yqi96/warranted --skill academic-writing --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Academic Writing?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Academic Writing
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/yqi96-academic-writing/badge)](https://www.skillsdirectory.com/skills/yqi96-academic-writing)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: academic-writing
description: Use when drafting or revising manuscript text that must project graph-backed scientific arguments into paper prose across Introduction, Related Work, Results, Discussion, Methods, figures, tables, source data, and citation handles.
---

## Load Required References

Use this skill as the manuscript-writing umbrella. Before drafting or revising a manuscript section, read the relevant reference file(s):

- `references/introduction-related-work.md` for Introduction, Background, Related Work, literature review, and literature-heavy Discussion passages.
- `references/results.md` for Results text, figure/table-linked result paragraphs, captions that state findings, and source-data references.
- `references/discussion.md` for Discussion, Conclusion, limitations, implications, and mixed literature/result synthesis.
- `references/methods-statistics.md` for Methods, Statistical Analysis, reproducibility, protocol, and analysis-authority prose.

Apply `literature-writing` whenever a passage uses published papers as evidence. `literature-writing` governs the literature evidence channel inside academic writing; this umbrella governs manuscript-level projection and cross-section consistency.

## Graph Translation Contract

Write the manuscript as a projection of the graph, not as a parallel argument.

Every load-bearing paragraph must have one paragraph proposition. A paragraph proposition may be new, reused, or revised, but it must be explicit before finalizing the paragraph. Transitional, roadmap, and signposting paragraphs do not need propositions; they must not introduce external facts, result interpretations, or synthesis claims.

Use the current Warranted contract: one Proposition node with `content`, `evidence`, `warrant`, `rebuttals`, and `qualifier`, stored in `.toulmin/graph.db`. Manuscript citations to graph-backed propositions use `\cite{prop_N}` handles. Do not use retired statement handles, node-type language, or retired completion gates.

Use this projection map:

| Manuscript unit | Graph source |
|---|---|
| Section thesis | section-level proposition, usually supported by paragraph propositions in `evidence.nodes` |
| Paragraph thesis | proposition content + qualifier |
| Reported paper finding | evidence proposition with source attachment + `\cite{prop_N}` |
| This-study result | evidence proposition with figure/table/source-data/method artifact |
| Reason why evidence matters | warrant text or promoted warrant proposition |
| Method/statistical authority | method/standard proposition cited from evidence or warrant |
| Limitation, disagreement, boundary condition | rebuttal proposition, usually created with `attacks={node, slot}` |

Do not write around a missing graph node. If the prose needs a scientific assertion, create or repair the graph first.

## Voice Separation

Keep four voices distinct:

- **Source voice**: what published work reports -> evidence proposition with source attachment.
- **Result voice**: what this study observed -> evidence proposition with data or figure artifact.
- **Inference voice**: what the evidence implies -> target proposition + warrant.
- **Judgment voice**: what the graph has earned -> qualifier after evidence and rebuttal assessment.

Do not use source or result voice to smuggle in inference. Sentences such as "Together, these findings show..." are proposition/warrant prose, not citation prose.

## Revision Protocol

When editing existing manuscript text:

1. Identify the paragraph proposition.
2. List every cited `prop_N`, figure, table, or source-data reference supporting it.
3. Check whether the paragraph's synthesis language matches the warrant.
4. If the text narrows, broadens, or changes the thesis, update the proposition or create a new one.
5. If an edit removes evidence, reassess the qualifier and inspect current warnings/findings.
6. If an edit introduces a limitation or contradiction, create or update a rebuttal proposition.

Text edits that change argument meaning are graph edits. Do not leave them as prose-only changes.

Files in this skill

  • SKILL.md3.5 KB
  • references/discussion.md2.1 KB
  • references/introduction-related-work.md1.9 KB
  • references/methods-statistics.md2.4 KB
  • references/results.md2 KB

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.