AI Agent Hub
Back to skills
SCI Literature Structured Reading icon

SCI Literature Structured Reading

Knowledge Management Updated 2026.08.30

Paste the following prompt into your AI chat to install this skill:

Install @user_4f291e59/shekang according to https://skillhub.cn/install/skillhub.md.

About this skill

Problem

Clinical guidelines often run to 50 or more pages. An abstract captures conclusions but hides recommendation strength, evidence grade, applicability conditions, and differences from prior or similar guidelines. This skill targets that kind of long, normative document and turns the original PDF into a structured, evidence-based reading.

How It Works

  • Input validation: It expects the original guideline PDF; titles, abstracts, or links alone stop the workflow.
  • Segmented full-text reading: It follows the table of contents to read methods, recommendations, and supporting evidence, reducing truncation and missed items.
  • Structured extraction: It extracts recommendation number, strength, evidence quality, and applicable population, then flags high-risk, workflow-changing, and immediately actionable items.
  • Comparative analysis: It searches earlier versions and related authoritative guidelines, then marks agreements, conflicts, additions, and deletions.
  • Personalized reporting: It combines the user's specialty to generate departmental implementation checks, resource needs, quality metrics, and evidence gaps, and can push an IMA note.

Boundaries

It suits clinical guidelines, expert consensuses, position statements, technical norms, and related evidence documents. It is not a substitute for summaries without full source. If comparison guidelines cannot be fully retrieved, the limitations are stated explicitly, and the output does not replace clinical judgment.

Use Cases

  • After receiving an 80-page WHO CVC guideline PDF, extract every recommendation, strength, evidence grade, and applicable population.
  • Before a unit QI review, compare the prior and current NCCN VTE guidelines item by item to identify added, removed, or changed recommendations.
  • While drafting a methods section, distinguish strong from conditional recommendations in a consensus and locate evidence-gap areas.
  • When curating department study materials, generate a structured guideline note with year, organization, and topic tags.

Best For

  • Clinicians: upload a guideline PDF and get a recommendation table, evidence grades, and actionable implementation checks.
  • Research assistants: compare consensuses and guidelines to locate evidence gaps and candidate study topics.
  • Quality improvement staff: extract quality metrics, safety boundaries, and workflow changes for infection or surgical quality projects.
  • Medical knowledge managers: archive long guidelines into structured notes tagged by year, organization, and topic.