Preface

Two common headaches when writing medical reviews: First, references are scattered across PubMed; manually exporting them one by one and checking Impact Factors is time-consuming and prone to omissions. Second, drafts assisted by AI often feature uniform sentences and lack the nuanced variations in subjective voice; editors can spot machine generation immediately, often resulting in revision comments consisting of a single line: “Language requires further modification.” Manual polishing involves spending excessive time on sentence structure and cohesion without establishing reusable rules.

The Academic Writing (De-AI-ification) introduced below integrates literature organization, drafting, AI trace removal, and proofreading into a single end-to-end pipeline. It also embeds authentic revision rules derived from actual sessions within the “De-AI-ification” and “Writing Enhancement” modules.

What is this?

Academic Writing (De-AI-ification) is an Agent Skill designed for medical review writing scenarios. Its coordinates are @user_a17d59be/academic-lit-review-writers. It is maintained by the community publisher user_a17d59be, with version 1.0.0, categorized as professional, and marked with a source of community.

It addresses the bridging issues among three tasks: batch retrieving literature from PubMed, filtering high-Impact Factor reviews for reference, and adapting Chinese review drafts to target journal paradigms for submission. The official description divides its capabilities into three groups: the Literature Pipeline (3 sub-capabilities), Drafting (4 sub-capabilities), and the Dual Engine for De-AI-ification and Writing Enhancement (7 sub-capabilities). The release notes further clarify that this pipeline is formed by the consolidation and enhancement of four capability sets: academic-lit-review-writer (PubMed retrieval → organization → filtering → drafting), humanizer-zh (removing AI traces), Chief Editor (editorial review), and academic-translation (Chinese-English comparison).

Core Functions

Listing only verified content from the materials:

  1. Literature Pipeline. Handles steps including batch PubMed retrieval, literature organization and bibliometrics, and screening reviews by Impact Factor. Trigger words provided in the release notes include: Organize literature, PubMed batch retrieval, bibliometrics, High-IF reviews, Impact Factor screening.
  2. Drafting. Handles the writing and submission polishing of Chinese medical journal reviews. Trigger words provided include: Write review, Review submission, Revise according to journal paradigm, Article structure optimization, Logical coherence, Style adaptation.
  3. Dual Engine for De-AI-ification and Writing Enhancement. Handles natural expression, with the goal of making machine-generated Chinese drafts read less like template outputs. Trigger words include: Remove AI flavor, Remove AI traces, Natural expression.
  4. Proofreading and Comparison. Chief Editor performs editorial review, while academic-translation provides Chinese-English comparison to facilitate checking terminology and expressions before submission.
  5. Built-in Revision Rules. The release notes mention that authentic revision rules precipitated from actual sessions are built into the De-AI-ification and Writing Enhancement modules. This means modification requirements accumulated during actual revisions are solidified into the Skill, rather than being explained ad-hoc every time.

Installation and Activation

The materials do not provide installation commands, and the official page could not be retrieved successfully (both https://www.skillhub.cn/skills/user_a17d59be/academic-lit-review-writers and https://api.skillhub.cn/user_a17d59be/academic-lit-review-writers returned retrieval failures). Therefore, no commands are provided in this article. The available verification information is as follows:

Coordinates: @user_a17d59be/academic-lit-review-writers
Name: Academic Writing (De-AI-ification)
Version: 1.0.0
Source: community
Category: professional
Publisher: user_a17d59be
Tags: requires_api_key=false

requires_api_key=false indicates that this Skill does not require API Key configuration, so no additional keys need to be prepared for activation. It is recommended to open the official page to confirm the current installation method before proceeding, as the distribution channels for community-published Skills may change.

Typical Usage

The trigger method is natural language. The trigger words provided in the materials cover the main stages from retrieval to polishing, allowing you to express a complete intent in a single sentence:

Organize Literature: Help me batch retrieve reviews in the XX field from PubMed, filter them by Impact Factor, and organize them into a list.

Write Review: Based on the high-IF reviews filtered above, write a Chinese medical journal review structured according to the target journal's paradigm.

Remove AI Traces: Revise this section to make it natural, remove the template tone, and smooth out the logical transitions.

Revise by Journal Paradigm: Adjust the heading hierarchy, abstract, and reference format according to the paradigm of XX journal.

Chinese-English Comparison: Create a Chinese-English comparison for the conclusion section; terminology needs to be unified.

It is recommended to proceed in stages: first run the Literature Pipeline to obtain the list and screening results, then proceed to Drafting, and finally use the De-AI-ification and Proofreading modules for processing. Compressing “Retrieval + Screening + Drafting + Polishing” into a single request makes it difficult to pinpoint which step caused issues during revision.

Applicable Scenarios and Notes

Suitable for three types of users: Researchers who need to perform bibliometrics and screening on PubMed without manual exports; Authors preparing Chinese medical journal review submissions who want structures to fit journal paradigms; Writers who already have AI drafts but are stuck on the “language doesn’t sound human” step.

Limitations to note:

  1. The materials do not specify whether this Skill depends on external database access permissions. Whether PubMed retrieval runs directly under the network conditions of the target environment needs to be verified by testing.
  2. requires_api_key=false only indicates that a key is not needed; it does not mean network access is not required.
  3. The built-in revision rules come from specific sessions. When encountering new journal paradigms or new disciplinary expression habits, supplementary explanations are still required.
  4. De-AI-ification is a rewriting method; it does not alter the data or conclusions of the paper and cannot be used to circumvent academic integrity requirements.
  5. The official page retrieval failed. Ratings, download counts, and other data are taken from pre-fetched materials. The installation method should be based on the official page.

Conclusion

The value of this Skill lies in stringing together the stages from literature organization to submission polishing. Specifically, it transforms De-AI-ification from a one-time manual rewrite into reusable, built-in rules, reducing repeated communication costs. As a community source with no API Key requirement, the barrier to trial is low.

Official Address: https://api.skillhub.cn/user_a17d59be/academic-lit-review-writers
SkillHub Page: https://www.skillhub.cn/skills/user_a17d59be/academic-lit-review-writers