MIT-licensed DeepSeek Harness plugin bundle that turns dsh into a read-only trading research workbench: typed market-data seam, deterministic indicators, interactive charts, and a strategy-integrity harness — with a hard no-execution boundary.
DSH integration
Native runtime
Author-claimed
Safety audit
Unaudited
Last verified
2026-09-06
License
MIT
01What can it help you accomplish?
Run read-only trading research on DeepSeek Harness — pull candles and compute a full multi-timeframe indicator regime, then chart it
Interactive candlestick cards (K-line + volume + SMA20/50/200) with chip-toggled indicator panes, plus a live chart column the user drives
Traders and analysts running DeepSeek Harness (dsh) who want research tooling without order execution
Audit a strategy's integrity — backtest fills against real candles, run a seeded baseline, check power
Verdicts (may honestly be NOT PROVEN), lookahead-leak lints, and a sample-size power check from the verdict harness
Quant researchers who need honest, testable strategy evaluation rather than a README 'research only' claim
Bring your own market data behind a typed seam
Swap the `market-data-provider` row to CSV, a broker API, ClickHouse, or CCXT — tools upstream never change
Engineers who want to plug their own market-data backend into dsh without rewriting tools
02How to install into DeepSeek Harness
Prerequisites
- DeepSeek Harness (`dsh`) installed — the plugin is a bundle of dsh plugins you stack on the stock `web` or `headless` profile
- A model key set via `DEEPSEEK_API_KEY` (or the Models page under `dsh web`)
Installation steps
- 01
Add the bundle: `dsh plugin --profile trading add @dsh-trading/bundle`
$ dsh plugin --profile trading add @dsh-trading/bundle
- 02
Add `@deepseek-ai/dsh-web-app` (or `@deepseek-ai/dsh-headless`) after `@deepseek-ai/dsh-base` in the profile's `dsh.profile.bundles` list
- 03
Configure a model key and run from a directory whose `./data` holds candles in the CSV layout
Verify the integration
- Verify the composed layers any time with `dsh --profile trading --dump-config`
03DSH integration and capability boundaries
Stacks on the stock `web` or `headless` profile as dsh plugins via `cordis.patch.yml` profile patches — no fork, no patched core.
Typed market-data seam (ctx.marketData)
any MarketDataProvider implementation→typed candle/symbol interface consumed by every tool
Deterministic multi-timeframe indicator regime
symbol + timeframe→`market_snapshot`: RSI, slow stochastic, ADX/DI, MACD, MFI, ATR, SMA/EMA, Bollinger with coarse state labels
Risk guard (read-only, no execution)
tool names at dsh's tools/pre-execute gate→refuses order-execution and fund-movement tool names
refuses execution-shaped tool names from any plugin at dsh's tools/pre-execute gateInteractive chart cards & live column
`market_snapshot` / `get_ohlcv` results from the model→interactive candlestick cards + a persistent, user-driven live chart column
04Who is it for? When not to use it?
Good for
- Traders and analysts running DeepSeek Harness (dsh) who want research tooling without order execution
- Quant researchers who need honest, testable strategy evaluation rather than a README 'research only' claim
- Engineers who want to plug their own market-data backend into dsh without rewriting tools
Not for
- No order-execution capability or execution seam exists; `@dsh-trading/risk-guard` refuses execution-shaped tool names at dsh's tools/pre-execute gate. Contributions adding live trading execution are out of scope. Nothing here is investment advice.
05Compatibility, maintenance and safety notes
- No order-execution capability or execution seam exists; `@dsh-trading/risk-guard` refuses execution-shaped tool names at dsh's tools/pre-execute gate. Contributions adding live trading execution are out of scope. Nothing here is investment advice.
- The chart-first shell (`@dsh-trading/client-frame`) is opt-in and deliberately absent from the bundle; you must disable dsh's `ui-layout` row and insert `@dsh-trading/client-frame` from your own profile patch.
- The `futu-api` SDK is a peer dependency deliberately unpinned — install the version matching your local OpenD (e.g. `10.9.x` SDK for a `10.9.x` OpenD); the provider warns at connect if the protocol versions differ.
MIT · research-only trading workbench; no tagged release (early scaffold)
06Frequently asked questions
How do I install dsh-trading on DeepSeek Harness?
Add the published bundle with `dsh plugin --profile trading add @dsh-trading/bundle`, then add `@deepseek-ai/dsh-web-app` (or `dsh-headless`) after `@deepseek-ai/dsh-base` in the profile's bundles list and set a model key. Verify with `dsh --profile trading --dump-config`.
Does dsh-trading place trades or give investment advice?
No. The project has no order-execution capability or execution seam, and risk-guard refuses execution-shaped tool names at dsh's tools/pre-execute gate. Nothing here is investment advice; contributions adding live trading execution are out of scope.
Can I use my own market data?
Yes. The `market-data` seam (`ctx.marketData`) is backed by a swappable `MarketDataProvider`. The bundle ships a CSV template (`provider-csv`); you repoint the `market-data-provider` row to a broker API, ClickHouse, or CCXT — tools upstream never change.
How are indicator values computed?
`market_snapshot` returns a whole multi-timeframe indicator regime (RSI, slow stochastic, ADX/DI, MACD, MFI, ATR, SMA/EMA, Bollinger) with coarse state labels. The math is pure and deterministic (textbook definitions, Wilder smoothing), so values reconcile against any charting platform.
07Related DSH workflows
weknora
by tencent
Open-source LLM knowledge platform: turn raw documents into a queryable RAG, an autonomous reasoning agent, and a self-maintaining Wiki.
honcho
by plastic-labs
Memory library for building stateful agents
mirage
by strukto-ai
The World's First Virtual Terminal for AI Agents
reme
by agentscope-ai
ReMe: Memory Management Kit for Agents - Remember Me, Refine Me.
08Data and sources
No fork, no patched core — just a bundle you stack on the stock `web` or `headless` profile.
built as plugins for [DeepSeek Harness](https://github.com/deepseek-ai/deepseek-harness) (`dsh`).
This page is generated from the project’s public documentation, repository metadata and a structured parse of DSH Plugins; last verified on 2026-09-06. Found an error? Submit a correction.
Best DeepSeek Harness Plugins
Twelve plugins worth installing first — picked from the whole catalog, across every category.