dsh-dataset-curation
Run the following command in DeepSeek Harness:
dsh plugin install satan9394/dsh-dataset-cution
Paste the following prompt into your AI chat to install this plugin:
Run dsh plugin install satan9394/dsh-dataset-curation in your terminal to install the plugin; the full source is available at https://github.com/satan9394/dsh-dataset-curation .
About this plugin
Dataset curation is often the unglamorous backbone of any LLM project: deduplication, filtering, format validation, quality spot-checks. Each step differs, teams scatter their scripts across repos, and standards exist only in Slack threads. dsh-dataset-curation pulls those scattered steps into a single structured workflow plugin for DeepSeek Harness, turning cleaning, QA, and annotation review into explicit, reproducible pipeline nodes.
As an agent-workflow plugin it does not replace your existing data pipeline. Instead it provides a staged scaffold inside the Harness workspace where you can attach your own scripts to pipeline nodes, or run built-in step templates for a full curation pass and then iterate. Input and output between nodes are recorded, making audits, rollbacks, and multi-person collaboration straightforward.
It is best suited for teams or solo practitioners preparing datasets for training, evaluation, or agent scenarios—especially when multiple people or multiple iterations need to share the same curation standard. If you already manage agent workflows in DeepSeek Harness, dsh-dataset-curation slots in as a dedicated stage, elevating data prep from an ad-hoc chore to a first-class part of your pipeline.
Use Cases
- Breaking down data cleaning and QA into composable pipeline steps for LLM training data preparation
- Replacing scattered one-off scripts with a single reproducible curation pipeline shared across the team
- Promoting dataset curation to a first-class, auditable stage inside a DeepSeek Harness agent workflow
Best For
- AI teams preparing training or evaluation datasets
- Developers managing agent workflows in DeepSeek Harness
- Solo practitioners needing reproducible and auditable curation pipelines
Related Plugins
A method pack that makes AI coding agents plan against your real baseline, prove completion with fresh evidence, and reduce reworks and unsafe changes.
Turns the DeepSeek Harness session into a captain that builds a durable sub-agent team, splits goals into dependency-aware tasks, and coordinates work via direct messages and a live Web UI.
Gives coding agents design judgment, letting Claude Code, Cursor, and 70+ agents generate and iterate high-quality UI, presentations, and graphics right from the terminal.
Run the Pi ecosystem's plugins on DeepSeek Harness, unmodified, via a compatibility layer that implements Pi's public extension ABI on DSH's native services.