← Back to Skills Marketplace
unisound-llm

unisound-literature-retrieval

by Unisound-LLM · GitHub ↗ · v1.0.0 · MIT-0
cross-platform ✓ Security Clean
96
Downloads
0
Stars
0
Active Installs
1
Versions
Install in OpenClaw
/install unisound-literature-retrieval
Description
医生端临床科研 — 医学文献问答与检索延展,在用户提供的题录/摘要片段上做 PICO 对齐与证据叙事。
README (SKILL.md)

医学文献检索(问答 / 延展)

概述

将「临床问题」与可选的本地已抓取题录或摘要片段一并交给内部医疗大模型,生成 PICO 对齐的问题重构、证据脉络梳理、检索式延展建议与知识空白提示。适用于开题、综述大纲、或向上级汇报前的文献叙事草稿。

本 skill 不直连 PubMed / CNKI 等外部索引passages 由上游爬虫、图书馆 API 或手工粘贴提供。与药企侧 literature-analysis 相比,本包更强调临床问题导向与「下一步可查什么」的延展,而非研发管线证据表。

业界脉络(写法参考,非功能承诺)

系统综述与快速证据合成中广泛采用 PICO(Population, Intervention, Comparator, Outcome) 框架约束检索问题;实践型产品常在「关键词布尔检索」之上叠加 语义段落排序引用级摘要。本 skill 在 prompt 层显式引导 PICO 与「可检索引言」,便于后续接入真实检索 API 时把同一 JSON 映射为查询 DSL。

快速开始

python3 scripts/run.py --input input.json --output output.json --appkey YOUR_KEY

输入字段(JSON)

字段 必填 说明
clinical_question 临床或科研问题(一句话或多句)
passages 文献片段数组:titleyearexcerpt(可无 PMID/DOI)
constraints 如人群、语种、研究设计偏好等

输出约定

{
  "skill": "医学文献检索",
  "status": "ok",
  "data": { },
  "text": "Markdown:PICO、要点、延展检索建议等"
}

参数

  • --input PATH必填
  • --output PATH:可选。
  • --appkey STRING必填。内部医疗大模型鉴权 key。

模型配置

  • https://maas-api.hivoice.cn/v1/chat/completions · u1-insuremed

边界

非实时文献库;不替代系统评价与人工全文阅读。若 passages 为空,模型仅基于通用医学知识给出检索策略级建议,须在呈现中区分「未锚定具体文献」。

Usage Guidance
This skill appears coherent and purpose-aligned. Before installing or using it, confirm you trust the publisher and API endpoint, use a protected appkey, and avoid sending patient-identifying or confidential clinical data unless your organization has approved that provider workflow.
Capability Analysis
Type: OpenClaw Skill Name: unisound-literature-retrieval Version: 1.0.0 The skill is a legitimate medical literature analysis tool designed to perform PICO (Population, Intervention, Comparison, Outcome) alignment using a Unisound medical LLM (hivoice.cn). The Python script (scripts/run.py) is well-structured, uses standard libraries for HTTP requests, and contains no evidence of malicious behavior, data exfiltration, or prompt injection risks.
Capability Assessment
Purpose & Capability
The stated purpose and code align: it reformats clinical literature questions into PICO-style evidence narratives and retrieval suggestions. The main user-visible consideration is that the provided question and passages are sent to an external/internal medical model endpoint.
Instruction Scope
The skill documents a manual command-line workflow with explicit input, output, and appkey arguments. There are no instructions to override user intent, run automatically in the background, or treat retrieved text as authoritative instructions.
Install Mechanism
There is no install spec or package dependency; the included script uses Python standard libraries. The internal _meta.json owner/slug differ from the registry listing, so users may want to confirm provenance or renaming history.
Credentials
The script reads a user-specified JSON file, optionally writes a user-specified output file, and calls the disclosed model API using a required appkey. This is proportionate to the skill purpose, but users should avoid including patient identifiers unless the provider relationship permits it.
Persistence & Privilege
No persistence, background service, privilege escalation, credential storage, broad local indexing, or autonomous account mutation is shown. The only file write is the optional user-specified output path.
How to Use
  1. Make sure OpenClaw is installed (local or Docker)
  2. Run the install command in chat: /install unisound-literature-retrieval
  3. After installation, invoke the skill by name or use /unisound-literature-retrieval
  4. Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
- Initial release of the 医学文献检索 (med-doctor-literature-retrieval) skill. - Supports clinical question entry and optional input of literature passages for PICO-aligned question restructuring, evidence synthesis, and retrieval suggestions. - Designed for clinician-facing research and reporting scenarios; does not connect directly to external databases (e.g., PubMed, CNKI). - Accepts structured JSON input and provides Markdown output with PICO breakdown, key points, and search extension advice. - Emphasizes clinical problem orientation and next-step search guidance over comprehensive evidence tables.
Metadata
Slug unisound-literature-retrieval
Version 1.0.0
License MIT-0
All-time Installs 0
Active Installs 0
Total Versions 1
Frequently Asked Questions

What is unisound-literature-retrieval?

医生端临床科研 — 医学文献问答与检索延展,在用户提供的题录/摘要片段上做 PICO 对齐与证据叙事。 It is an AI Agent Skill for Claude Code / OpenClaw, with 96 downloads so far.

How do I install unisound-literature-retrieval?

Run "/install unisound-literature-retrieval" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is unisound-literature-retrieval free?

Yes, unisound-literature-retrieval is completely free, licensed under MIT-0. You can download, install and use it at no cost.

Which platforms does unisound-literature-retrieval support?

unisound-literature-retrieval is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created unisound-literature-retrieval?

It is built and maintained by Unisound-LLM (@unisound-llm); the current version is v1.0.0.

💬 Comments