dsh-academic-research-skills
文件与数据 活跃维护

dsh-academic-research-skills

nullptr-DZF/dsh-academic-research-skills

集成Claude学术研究技能模块至工具生态,支持文献检索、论文结构梳理、研究问题拆解等核心能力,无需额外配置即可在工作流中直接调用,降低学术研究工具使用门槛。

5
Stars 标星
0
Forks 分支
5
Watchers 关注
0
Open Issues
JavaScript
主要语言
NOASSERTION
开源协议
1.1 MB
仓库大小
25 天前
最后推送
一键安装扩展 / 插件指令
dsh plugin --profile web add github:nullptr-DZF/dsh-academic-research-skills
git clone https://github.com/nullptr-DZF/dsh-academic-research-skills.git
git clone git@github.com:nullptr-DZF/dsh-academic-research-skills.git
README.md main

dsh-academic-research-skills

English · 中文

DeepSeek Harness(DSH)插件:把 Claude Code 的
Academic Research Skills (ARS)
(v3.21.1,CC-BY-NC-4.0)移植为 DSH 技能。

安装后自动获得:

  • 4 个模型可调用技能deep-research(13-agent 调研团队,8 种模式)、
    academic-paper(12-agent 论文写作,11 种模式)、academic-paper-reviewer
    (5 席位同行评审面板)、academic-pipeline(10 阶段全流水线编排)
  • 16 个用户可调用命令/ars-full/ars-plan/ars-reviewer
    /ars-lit-review/ars-abstract/ars-revision/ars-outline
    /ars-citation-check/ars-disclosure/ars-format-convert
    /ars-rebuttal-audit/ars-revision-coach/ars-cache-invalidate
    /ars-mark-read/ars-unmark-read/ars-3w

署名 / Attribution:本项目为 ARS 的 DSH 移植,技能内容与命令衍生自
Cheng-I Wu(Imbad0202)的
academic-research-skills
全部内容(含移植层)遵循 CC-BY-NC-4.0(非商业用途),详见
NOTICE.mdLICENSE

安装

要求:DeepSeek Harness(dsh)0.1.x,pnpm 可用。

# 方式一:GitHub 仓库(公开后可用)
dsh plugin --profile web add github:nullptr-DZF/dsh-academic-research-skills

# 方式二:本地路径(开发调试)
dsh plugin --profile web add <本仓库绝对路径>

# 重启 dsh web 后生效(技能目录会自动出现)

安装后 dsh.profile.bundles 会追加 dsh-academic-research-skills,启动时
lib/startup.js 把 20 个技能注册进 ctx.skills

使用

用法 操作
模型自动路由 直接说“做学术研究 / 文献综述 / 帮我审稿 / 写论文”等(技能描述带触发词)
技能注入 输入 /deep-research/academic-pipeline 等技能名
斜杠命令 输入 /ars-full(全流程)、/ars-reviewer(审稿)、/ars-plan(规划)等

与上游的差异

  • hooks 不移植:上游 hooks/hooks.json(写保护 guard、SessionStart 播报)
    是 Claude Code 专属;DSH 无 hook 系统,写保护由 DSH 自身文件沙箱承担。
  • 确定性脚本需上游仓库:引用核验 gate 等 scripts/*.py 不在本包内;
    需要时 clone 上游仓库并按其 README 安装 Python 依赖(pypdf 等)。
  • 子代理:上游 39 个提示角色默认内联运行;3 个插件级 agent 文件随技能包
    分发,需要时可把其内容作为 subagent 工具的 prompt 使用。

重新生成命令包装

commands/ 由上游 commands/ars-*.md 转换而来(加入
disable-model-invocation: true + user-invocable: true):

node scripts/regenerate-commands.mjs <上游仓库路径>

English

A DeepSeek Harness plugin
port of Claude Code's
Academic Research Skills (ARS)
(v3.21.1, CC-BY-NC-4.0).

Provides, out of the box:

  • 4 model-invocable skills: deep-research, academic-paper,
    academic-paper-reviewer, academic-pipeline
  • 16 user-invocable commands: /ars-full, /ars-plan, /ars-reviewer,
    /ars-lit-review, /ars-abstract, /ars-revision, /ars-outline,
    /ars-citation-check, /ars-disclosure, /ars-format-convert,
    /ars-rebuttal-audit, /ars-revision-coach, /ars-cache-invalidate,
    /ars-mark-read, /ars-unmark-read, /ars-3w

Attribution: this is a DSH port of ARS; skill content and commands are
derived from Cheng-I Wu's
academic-research-skills.
Everything in this repo (including the porting layer) is licensed
CC-BY-NC-4.0 — non-commercial use only. See NOTICE.md and
LICENSE.

Install

Requires DeepSeek Harness 0.1.x and pnpm.

dsh plugin --profile web add github:nullptr-DZF/dsh-academic-research-skills
# or from a local checkout:
dsh plugin --profile web add /path/to/this/repo
# restart `dsh web`; the skills then appear in the session catalog

The bundle patch mounts lib/startup.js, which registers the 20 skills into
ctx.skills at boot.

Differences from upstream

  • Hooks are not ported: upstream hooks/ are Claude Code specific; DSH has
    no hook system (write protection is DSH's file sandbox).
  • Deterministic scripts live upstream: scripts/*.py (citation
    verification gate, etc.) are not bundled; clone the upstream repo if you need
    them.
  • Subagents: the 39 prompt roles run inline by default; the 3 plugin-level
    agent files ship with the skill bundles and can be used as subagent prompts.

Regenerating command wrappers

commands/ is generated from upstream commands/ars-*.md:

node scripts/regenerate-commands.mjs <upstream-checkout-path>