← 返回 Skills 市场

unisound-paper-writing-assist

Name: unisound-paper-writing-assist
Author: unisound-llm

作者 Unisound-LLM · GitHub ↗ · v1.0.0 · MIT-0

cross-platform ✓ 安全检测通过

总下载

当前安装

版本数

在 OpenClaw 中安装

/install unisound-paper-writing-assist

功能描述

医生端临床科研 — 论文写作辅助，按 IMRaD 等结构将要点扩写为学术中文/英文段落草稿。

使用说明 (SKILL.md)

论文写作辅助

概述

根据作者提供的章节类型与要点列表（可含已测主要结局、样本量、统计方法名等），由内部医疗大模型生成 IMRaD 风格段落草稿（引言 / 方法 / 结果 / 讨论 / 摘要等），并附带报告规范意识层面的检查清单式提示（如 CONSORT / STROBE 等仅为写作提醒，不做合规认证）。

本 skill 不代写未披露数据，不生成伪造统计结果；数值须由作者在 notes 中显式给出，模型只做语言组织与逻辑衔接。

业界脉络（写法参考，非功能承诺）

医学期刊写作长期以 IMRaD 结构为主流；观察性与干预性研究分别有 STROBE、CONSORT 等报告规范用于提高可重复性。生成式辅助写作的常见产品形态为「大纲 → 段落 → 术语一致性润色」流水线。本 skill 将「章节 + 要点」映射为该流水线的最小可调用单元，便于与编辑器或投稿前校验工具组合。

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
`section`	是	如 `abstract`、`introduction`、`methods`、`results`、`discussion`
`notes`	是	字符串数组：每点一条事实或待写要点
`language`	否	`zh`（默认）或 `en`
`journal_style_hint`	否	目标期刊或字数上限等自由文本

输出约定

{
  "skill": "论文写作辅助",
  "status": "ok",
  "data": { },
  "text": "Markdown：段落草稿 + 规范提醒"
}

参数

--input PATH：必填。
--output PATH：可选。
--appkey STRING：必填。

模型配置

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

边界

草稿须由作者修改、署名并对科学内容负责；禁止将模型输出作为未经验证的结论直接投稿。

安全使用建议

Before installing, verify the publisher identity, use a dedicated API key, and only provide research notes that are approved for transmission to the hivoice LLM API. Do not include patient-identifying or confidential clinical information unless your organization permits it.

功能分析

Type: OpenClaw Skill Name: unisound-paper-writing-assist Version: 1.0.0 The skill is a legitimate medical paper writing assistant that facilitates the generation of academic drafts using a specific LLM API (maas-api.hivoice.cn). The Python script (scripts/run.py) is a clean implementation using standard libraries to handle HTTP requests and JSON processing, with no evidence of malicious intent, data exfiltration, or unauthorized execution.

能力评估

ℹ Purpose & Capability

The SKILL.md and code consistently describe converting user-provided section notes into IMRaD-style manuscript draft text. The medical-writing use is purpose-aligned, but outputs still require human scientific review.

✓ Instruction Scope

The system instructions explicitly prohibit inventing data, sample sizes, p-values, or references and require missing information to be listed rather than fabricated.

ℹ Install Mechanism

There is no install script or dependency installation; the runnable script uses Python standard libraries. However, the packaged _meta.json owner/slug differ from the registry metadata, so publisher/provenance should be checked.

ℹ Credentials

The script reads a user-specified JSON input file, posts its contents to the disclosed hivoice LLM API, and writes an optional user-specified output file. This is proportionate to the stated purpose, but data leaves the local environment.

ℹ Persistence & Privilege

No background persistence, elevated local privileges, or destructive actions are shown. The skill does require a user-supplied app key for the external model API.

如何使用

确保已安装 OpenClaw（本地或 Docker 部署）
在对话框中输入安装命令：/install unisound-paper-writing-assist
安装完成后，直接呼叫该 Skill 的名称或使用 /unisound-paper-writing-assist 触发
根据 Skill 的参数说明提供必要输入，即可获得结构化输出

版本历史

v1.0.0

- Initial release of the skill, now named "med-doctor-paper-writing-assist". - Generates IMRaD-structured academic drafts in Chinese or English from user-provided key points. - Includes writing reminders aligned with major reporting guidelines such as CONSORT and STROBE. - Does not fabricate results; authors must supply all data for model to organize and draft. - Designed for integration with manuscript editors and pre-submission checking tools.

元数据

Slug unisound-paper-writing-assist

版本 1.0.0

许可证 MIT-0

累计安装 0

当前安装数 0

历史版本数 1

常见问题