Academic Writing Harness
Paste the following prompt into your AI chat to install this skill:
Please install @user_d0deb03a/academic-writing-harness according to https://skillhub.cn/install/skillhub.md.
About this skill
Problem
The recurring issue in academic writing is not surface wording, but argument gaps: skipped inferences, weak evidence, and reviewer comments that are hard to decompose. For experimental papers, reviews, and grant proposals, authors often receive feedback like “logic does not follow” or “evidence is insufficient” without a repeatable audit process.
How It Works
Academic Writing Harness turns review into executable prompts: use the reverse-outline prompt to infer premises from the conclusion and locate gaps; use the four-dimensional evidence-chain prompt to test whether materials support claims; and use reviewer-response plus minimal-repair prompts for controlled fixes. It follows “continuous generation + deterministic validation + minimal counterexample repair,” prioritizing P0, P1, and P2 risk levels across iterations.
Boundaries
This approach fits experimental papers, literature reviews, and grant proposals best. For purely theoretical or concept-driven work, authors need domain anchors. Its output is an evidence-strength assessment, not a binary verdict, so final quality still depends on the author’s judgment of novelty, clarity, and format compliance.
Use Cases
- When revising an experimental paper draft, use the reverse-outline prompt to infer premises from conclusions and locate skipped inferences or unstated assumptions.
- After receiving a reviewer comment about weak logic, use the minimal-repair prompt to reduce the issue to a smallest counterexample and edit only necessary sentences.
- While drafting a grant proposal or research plan, use the evidence-chain evaluation prompt to check whether data, methods, and conclusions support the stated objectives.
- Before submission, run a dual-path review so the model generates both critique and defense to expose weak assumptions in the paper's position.
Best For
- Graduate students revising experimental SCI papers and needing reviewer comments broken into locatable edits
- Postdoctoral researchers evaluating grant technical routes and checking whether claims support the stated objectives
- Early-career researchers preparing literature reviews or research plans and needing to check evidence chains and conclusion boundaries
- Advisors or academic editors guiding paper writing and needing a repeatable review checklist for students
Related Skills
A nanoAR projection advisor that applies fixed optics formulas to select films, projectors, and quote templates.
Run the ORI Nextflow collection on the Tencent Health Genomics Platform for protein generation, signal peptide, solubility, thermostability, and structure prediction.
Evaluates regional resources, industrial base, and innovation capacity to recommend future industry directions with actionable layout guidance.
A pre-IPO tax and internal-control framework covering three-year cleanup, restructuring tax, controls, remediation, related-party transactions, and compliance reporting.