doubao-academic-evaluator

Evaluates research ideas and reviews paper drafts with severity-ranked feedback.

2|Updated Aug 9, 2026
One-click install
npx skills add https://github.com/DeepJH/doubao-skill-and-info --skill doubao-academic-evaluator-deepjh
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: doubao-academic-evaluator
Source: https://github.com/DeepJH/doubao-skill-and-info/tree/main/skills/doubao-academic-evaluator
Command: npx skills add https://github.com/DeepJH/doubao-skill-and-info --skill doubao-academic-evaluator-deepjh

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Researchers often lack an honest, experienced reviewer before committing months to an idea or submitting a paper. This Skill acts as a senior professor who diagnoses research ideas and pre-submission drafts, identifying fatal flaws and giving clear go/no-go verdicts without rewriting the work itself. ## Core Features & Use Cases - Idea Evaluation: Scores a research idea across five dimensions (accuracy, speed, robustness, cost, breadth), checks ten classes of fatal flaws, verifies novelty via scholar_search, and concludes with "worth doing / revise first / change direction". - Pre-Submission Review: Reviews a finished draft layer by layer (logic chain, claim-evidence alignment, novelty, language, AI-tone patterns, figures, LaTeX formatting), ranking every issue as critical, major, or minor with exact locations in the text. - Cross-Disciplinary Frameworks: Routes non-STEM work (humanities, social science, finance/economics, law) to domain-appropriate evaluation dimensions and fatal-flaw lists instead of applying STEM criteria. - Use Case: Before submitting to a conference, paste your draft and ask for a review; you receive a prioritized issue list with quoted evidence, concrete fixes, and a clear submission recommendation. ## Quick Start Ask the assistant to review your attached paper draft before submission and list the critical issues that could cause rejection.

Frequently Asked Questions about doubao-academic-evaluator

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I evaluate whether a research idea is worth pursuing?▼

Describe your idea and ask for an evaluation. The Skill restates the contribution, checks ten classes of fatal flaws, scores five dimensions from a default of 5 with evidence required for higher scores, and concludes with worth doing, revise first, or change direction.

How to review a paper draft before conference submission?▼

Provide the draft and ask for a pre-submission review. The Skill checks the introduction argument chain, claim-evidence alignment, novelty via scholar_search, language issues, AI-tone words, figures, and LaTeX formatting, ranking each issue as critical, major, or minor with exact locations.

Does the evaluation work for humanities or social science research?▼

Yes. The Skill detects the research paradigm by method, not discipline name, and switches to domain-specific frameworks for humanities, empirical social science, finance/economics, and law, each with its own scoring dimensions and fatal-flaw lists.

Can it rewrite or polish my paper after the review?▼

No. This Skill only diagnoses and recommends; it never rewrites text or creates figures. When rewriting is needed, it points you to the companion skill doubao-academic-polish for drafting and polishing work.

How does it verify novelty without inventing citations?▼

It uses scholar_search with 2-3 keyword sets derived from your idea, compares the 3-5 closest published works along explicit difference axes, and caps total cited references at 15. Search metadata is never used to fabricate experimental numbers.