DeepScient plugin
A DeepSeek Harness plugin links research stages to data, experiments and checks before the next stage begins.
- 865 ★ on the repository
- Last commit
- MIT
Extended access
Why this level
- The plugin writes research files and connects tools and sources; model outputs and third-party skills require review.
Install
Manual install
plugin_manager install_bundle <仓库根>\dsh-scienceThe Quick Start command runs inside a dsh session. Substitute the absolute path to the downloaded repository root.
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- The plugin writes research files and connects tools and sources; model outputs and third-party skills require review.
This is third-party code. Review the repository files before installing.
What it does
DeepScient adds scientific tools and workflows to DeepSeek Harness. The agent helps move from a question and literature search through experiments, analysis, writing and reviewer responses. It checks recorded project state and links conclusions to evidence. The dsh-science core can be installed separately from the DeepScient shell. Human judgment remains responsible for ethics, novelty and construct validity; the agent prepares material for review.
Who it is for. Research groups using DeepSeek Harness to keep experiments, findings and evidence in one project.
Good fit when
- A literature review must lead into an experiment tied to the research question.
- Analysis needs recorded environments, versions and random seeds.
- Before drafting a paper, you check which claims have supporting data.
- Reviewer responses require revisiting experiments and project history.
Not a fit when
- You expect automatic ethical approval of a study.
- A report is expected without data or reproducible experiments.
- No compatible DeepSeek Harness is available for the research core.
Example request
Connect dsh-science to a test profile and help prepare an experiment plan without inventing results.Limitations
The research-core plugin requires DeepSeek Harness; packaged desktop apps belong to a separate shell distribution. The documented source Quick Start was checked on Windows 10/11 and needs Node.js 22.5 or later, so do not assume those commands work unchanged elsewhere. Extra research skills download third-party repository content; inspect the installation plan first. MCP tool registration has been checked, but the author notes that a full model-driven PubMed search round trip remains unverified.
How to disable. Disable the dsh-science bundle in the selected Harness profile’s Plugins page and stop the research task. Preserve project materials separately.
FAQ
Is the DeepScient shell required?
No. The dsh-science package can run separately inside dsh.
How do I install without a command?
Select the local dsh-science folder on the DeepSeek Harness Plugins page.
How do I preview extra skills?
Run `node deepscient/scripts/install-skills.mjs --plan` from the repository.
Why is the agent preset missing in headless mode?
When you start without the interface, the agent preset is not loaded. Use the web interface to check which skills the session has loaded.
Which literature sources are available?
Sources include Crossref, OpenAlex and arXiv. Select those relevant to the topic and verify access.
Discussion
No questions or reviews yet.
Related
PluginsWith Deep Research, the agent gathers scientific papers and data, compares sources and prepares a PDF or DOCX report.
A Claude Code skill for scientific literature reviews: an interview, Scopus and OpenAlex search, DOI checks and a final docx file
A marketplace of five Claude Code plugins for researchers: literature search, research design, project memory and manuscript work
A research loop that keeps evidence: a literature review, paper review or survey runs in steps with checks and resume.
A library of 400+ academic skills: literature search, reviews, statistics, scientific writing and grant proposals
A plugin helps agents maintain project documentation, decisions and verification evidence across development sessions
README in short
The shell and research core have separate versions. The project interface includes stages, topic storage, connections and model-use accounting.