dsh-quant
dsh-quant is a DeepSeek Harness plugin package (npm: dsh-quant) that registers 59 quant_* tools across six domains — data, alpha, ML, risk, execution, and ecosystem — and ships a one-call PDAT→PET research pipeline. It is written in TypeScript with zero runtime dependencies, includes a zero-dependency Node CLI, and pairs with a separate dsh-quant-ui workbench repository.
Installation
No reliable install command was detected. Check the project README for installation instructions.
Overview
dsh-quant is a DeepSeek Harness plugin package (npm: dsh-quant) that registers 59 quant_* tools across six domains — data, alpha, ML, risk, execution, and ecosystem — and ships a one-call PDAT→PET research pipeline. It is written in TypeScript with zero runtime dependencies, includes a zero-dependency Node CLI, and pairs with a separate dsh-quant-ui workbench repository.
- Registers 59 quant_* tools across six domains: data, alpha, ML, risk, execution, and ecosystem.
- quant_research_pipeline runs the full PDAT→PET chain in one call, returning candles, metrics, risk, fund, and report.
- Organized as five pluggable modules: dsh-data, dsh-alpha, dsh-ml, dsh-risk, and dsh-execution.
- Install with npm i dsh-quant and add one line to cordis.yml; the tools auto-register.
- Zero runtime dependencies and 21 TypeScript source files; README cites 298 tests.
- Numerical methods are pure functions with hand-computed baselines, per the README's AI-native design statement.
- Outputs use equal-length null alignment and canonical JSON, with render separated for human-readable prose.
- Ships a zero-dependency Node CLI (repo, history, kline, browse) plus a separate dsh-quant-ui workbench repo.