AI & Skills
Overview
Deepcoin Open API documentation can be used directly as AI context, and can also be organized as reusable agent skills. This page combines the "AI documentation" approach and the "skill-based routing" approach so AI tools can better understand Deepcoin market, account, and trading APIs.
This design is based on two ideas:
- Provide Markdown documentation that an AI tool can read directly
- Provide focused
SKILL.mdfiles so an agent can route requests by intent
Skill Architecture
The current design splits Deepcoin capabilities into three skills:
| Skill | Scope | Recommended Use |
|---|---|---|
deepcoin-market | Public REST and public WebSocket market data | Quotes, order book, trades, candles, product info, funding rate |
deepcoin-account | Authenticated account and private WebSocket state | Balance, positions, bills, UID, leverage, sub-account, internal transfer |
deepcoin-trade | Authenticated trading and strategy workflows | Place, cancel, replace, trigger, TP/SL, batch, copy trade, strategy order |
Skill files are stored in this repository:
skills/deepcoin-market/SKILL.mdskills/deepcoin-account/SKILL.mdskills/deepcoin-trade/SKILL.md
Design Principles
1. Route by user intent
The same symbol such as BTC-USDT-SWAP may appear in market, account, and trading requests. Skills should route by user intent instead of by symbol:
- "What is the latest BTC funding rate?" ->
deepcoin-market - "Show my BTC-USDT-SWAP position" ->
deepcoin-account - "Place a BTC-USDT-SWAP long order" ->
deepcoin-trade
2. Treat private APIs as signed APIs
Any skill that touches private REST endpoints must follow the signature rules in authentication:
DC-ACCESS-KEYDC-ACCESS-SIGNDC-ACCESS-TIMESTAMPDC-ACCESS-PASSPHRASE
3. Prefer code generation before execution
At this stage, Deepcoin skills are designed primarily as documentation-centric skills. They help AI tools:
- answer API questions accurately
- generate REST/WebSocket code samples
- compose correct request bodies
- explain signing, symbols, and field semantics
If Deepcoin later provides an official CLI or MCP server, the same skill structure can be extended from "documentation skill" to "executable skill".
Suggested Skill Packaging
For an external AI tool or code agent, package the skills with this structure:
deepcoin-skills/
skills/
deepcoin-market/
SKILL.md
deepcoin-account/
SKILL.md
deepcoin-trade/
SKILL.md
docs/
authentication.md
DeepCoinMarket/
DeepCoinAccount/
DeepCoinTrade/
publicWS/
privateWS/
The SKILL.md files define when to use each skill, while the Deepcoin docs remain the authoritative reference for endpoints and fields.
Recommended Next Step
Read Usage Guide for installation patterns, slash-command examples, and natural-language invocation guidance.