> For the complete documentation index, see [llms.txt](https://smol404.gitbook.io/smol404-docs/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://smol404.gitbook.io/smol404-docs/launchpad/agents.md).

# Agents

Each user authenticates in Telegram with an access code that will be shared on the dapp. All messages after login are interpreted in natural language and routed to the matching handler.

<figure><img src="/files/zDCWQX8lMIBxKRosEs53" alt=""><figcaption></figcaption></figure>

#### Type 1 - ECHO

Echo is the simplest agent. It takes the user’s message and writes it on-chain through the Echo contract. There is no token routing, no vault debit, and no trading logic. It is intended for on-chain logging, announcements, or any payload that must be timestamped on BSC.

e.g.\
Hello World\
The bot replies with the transaction hash and a BscScan link.

#### Type 2 - TRADER

Trader is the core on-chain execution agent. It buys and sells tokens against the agent vault, checks balances, reports prices, and can watch a target price before firing a trade.

Supported actions: instant buy/sell, balance check, price query, limit buy/sell, cancel and list limit orders. Amounts are expressed in BNB. The token may be a known symbol (CAKE, BTC) or a 0x address.

e.g.\
Buy 0.001 BNB of CAKE\
Check balance of CAKE\
Buy 0.0001 BNB of BTC when BTC price is 68260\
Execution always goes through handleAction on the Trader contract. The technical wallet pays gas; the vault of that tokenId pays the trade. If the vault is empty, the transaction reverts even if the operator wallet still holds BNB.

#### Type 3 - DCA

DCA uses the same Trader contract and the same vault as Type 2, but on a schedule. The user starts a recurring buy or sell; the backend runs an async loop and submits buy\_token or sell\_token at the given interval. Several DCAs can run in parallel for the same user.

e.g. Start DCA buy BTC with 0.0001 every 5 minutes Stop DCA List DCAs Jobs live in process memory. A bot restart stops all schedules; the user must start them again.

#### Type 4 - NEXUS (CEX trader)

Nexus is the only agent that does not use the technical wallet or the on-chain vault. It trades Spot and Futures on Binance through Agentic MCP, on the user’s Agentic sub-account.

Authentication is required before any order:

The user sends the code, then connect binance.\
The bot returns a Binance OAuth link bound to that Telegram chat.\
After authorization, the callback stores an access token against the chat\_id.\
Later MCP calls are sent with Authorization: Bearer \<token>.

\
e.g.\
connect binance\
Buy 10 USDT of BTC spot\
Short 0.01 ETH futures\
Show my Binance balances\
disconnect binance\
Nexus can list MCP tools, read balances, and place spot or futures orders. It never signs a BSC transaction.

#### Type 5 - ECHO HYBRID (Arbitrage)

cho Hybrid needs no Binance account. It reads public CEX order books (MCP, with REST fallback), compares the CEX mid price to the on-chain reference price, and, if the spread is above the threshold (default 0.4%), executes one on-chain leg through the Trader vault — the same path as Type 2.

CEX mid above on-chain → buy on-chain (asset cheaper on DEX).\
CEX mid below on-chain → sell on-chain.\
It does not place a CEX order. The token must resolve in tokens.json (or as an address) and must have an on-chain price feed.

e.g.\
Scan arb for CAKE\
Scan arb for CAKE min spread 0.5%

#### Type 6 - SPECTER (Copy-trading)

Specter watches an external wallet via Binance Skills Hub tracker tools. When a buy or sell is detected, it mirrors the side on-chain with a fixed BNB size taken from the agent vault. Polling interval is at least 30 seconds.

e.g.\
Copy wallet 0xabc... with 0.01 BNB every 60s\
Inspect wallet 0xabc...\
List copies\
Stop copy\
Copy jobs are in-memory. A restart stops mirroring. Transfers without a clear buy/sell side are ignored.

#### Type 7 - ORACLE (Smart DCA)

Oracle is a buy-only DCA whose size is adjusted before every cycle. It pulls funding and market sentiment from MCP / Skills Hub, then asks the model for a multiplier between 0.25× and 2.0× of the base amount (smaller when the market looks overheated, larger when it looks fearful). The sized order is then executed on-chain through the same vault as Types 2 and 3.

If modulation fails, the base amount is used.

e.g.\
Start smart DCA buy BTC with 0.01 every 10 minutes\
List smart DCAs\
Stop smart DCA\
Summary


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:

```
GET https://smol404.gitbook.io/smol404-docs/launchpad/agents.md?ask=<question>&goal=<endgoal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.
