メニュー
Conducts multi-dimensional manuscript review. Use when evaluating draft quality across argumentation, clarity, evidence, structure, and academic style before submission.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
| name | code-review-and-quality |
| description | Conducts multi-dimensional manuscript review. Use when evaluating draft quality across argumentation, clarity, evidence, structure, and academic style before submission. |
Multi-dimensional manuscript review with quality gates. Every draft section gets reviewed before inclusion in the final paper. Review covers five axes: argumentation, clarity, evidence, structure, and academic style.
The approval standard: A section is ready when it clearly advances the paper's argument, is supported by evidence, and meets venue standards — even if it could still be polished further. Perfect prose doesn't exist; the goal is continuous improvement toward submission readiness.
When NOT to use: During initial brainstorming or topic exploration where rough notes are expected.
Does the paper present a sound, well-reasoned argument?
Can another researcher understand this without the author explaining it?
Is every claim backed by verifiable evidence?
Does the paper follow a logical progression?
Does the writing meet academic publishing standards?
| Severity | Meaning | Action Required |
|---|---|---|
| Critical | Blocks submission | Logical fallacy, fabricated citation, missing key section |
| Important | Should fix | Weak argument, missing related work, unclear methodology |
| Suggestion | Nice to have | Better word choice, additional analysis, formatting polish |
| Rationalization | Reality |
|---|---|
| "The reviewer will understand what I mean" | If it's ambiguous to you, it's ambiguous to them. Be explicit. |
| "I'll add citations later" | Unsupported claims become habits. Cite as you write. |
| "The results speak for themselves" | Results need interpretation. Explain what they mean and why they matter. |
| "This limitation is obvious, no need to state it" | Reviewers respect transparency. Unstated limitations look like blind spots. |
| "It's just a first draft" | First drafts set the structure. Review early to avoid compounding problems. |
After review is complete:
docs/research/reviews/Records research methodology decisions and documentation. Use when making structural decisions about the paper, choosing between analysis approaches, selecting venues, or when you need to record rationale that collaborators and future reviewers will need to understand the research choices.
Breaks research paper work into ordered tasks. Use when you have a research topic or outline and need to break the writing into manageable stages. Use when the paper feels too large to start, when you need to estimate scope, or when sections can be drafted in parallel.
Handles final assembly, formatting, and delivery of research papers. Use when compiling a manuscript for submission, formatting for a target venue, running pre-submission checklists, or packaging supplementary materials.
Defines and tracks the complete end-to-end research paper lifecycle — from idea inception to published paper. Enforces readiness gates at every phase transition so the agent always reports current status and what's needed to advance. Use when starting a new paper, assessing progress, or deciding what to work on next.