← 返回 Skills 市场
happyrstudent

AI News BNB Trader

作者 happyRstudent · GitHub ↗ · v0.1.0
cross-platform ⚠ suspicious
283
总下载
0
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install ai-news-bnb-trader
功能描述
TypeScript (Node.js 20+) AI news-driven BNB strategy trading bot for BSC. Use when user wants event/news sentiment signals, strict risk controls, and automat...
使用说明 (SKILL.md)

AI News BNB Trader

Run an event-driven BSC trading bot using news sentiment + risk gates.

Safety first

  • Default DRY_RUN=true.
  • Never print private key/seed in logs.
  • Only whitelist assets (WBNB, USDT/BUSD/USDC).
  • Panic mode immediately disables trading.

Commands

npm run start -- start
npm run start -- status
npm run start -- panic
npm run start -- revoke-approvals
npm run key:encrypt -- --out ./secrets/key.json

News modes

  • Poll REST via NEWS_API_URL every NEWS_POLL_SECONDS
  • Optional WebSocket via NEWS_WS_URL

Both modes dedupe on news.id and log failures with exponential backoff.

Signal models

  • RuleSignalModel (default): keyword rules with explainable reasons
  • OpenAISignalModel (optional): enabled when OPENAI_API_KEY is set; timeout + fallback to rules

Strategy

  • Buy WBNB with stablecoin if sentiment * impact >= BUY_THRESHOLD and confidence >= MIN_CONF
  • Sell WBNB to stablecoin if sentiment * impact \x3C= -SELL_THRESHOLD and confidence >= MIN_CONF
  • Enforce all risk gates before execution

Risk controls

  • Max order notional / max position pct / daily trade cap
  • Daily loss cap (MTM approximation)
  • TP/SL from avg entry
  • Cooldown between trades
  • Slippage cap from quote
  • Consecutive failures -> SAFE_MODE

Notes

  • Prefer private RPC (PRIVATE_RPC_URL) when provided.
  • For production: validate token/DEX addresses and add deeper MEV defenses.
安全使用建议
This repository contains a real trading bot that will perform on-chain swaps and can revoke approvals — it legitimately needs a wallet private key (or an encrypted key + passphrase), RPC endpoints, and optional API keys (OpenAI, 1inch). The registry metadata omits these required env vars, which is a red flag: before running, review the code locally, and follow these precautions: - Do not supply your main/large-balance wallet private key. Use a fresh wallet with minimal funds for testing. - Keep DRY_RUN=true while evaluating; verify behavior in dry-run mode and inspect logs and saved state-data before enabling real trades. - Inspect NEWS_API_URL and NEWS_WS_URL sources to ensure they are trusted (the bot executes trades based on that feed). - If using the encrypt CLI, the code uses AES-256-GCM + scrypt; verify encryption/decryption locally and ensure KEY_PASSPHRASE is handled securely. Note the logger redacts some envs but not KEY_PASSPHRASE or RPC secrets — avoid putting passphrases into logs or shared environments. - Audit and limit RPC and DEX API credentials; prefer a private RPC provider and restrict API keys to minimal scopes where possible. - Consider running in an isolated environment (container, VM) and reviewing the package.json dependencies with npm audit before install. Because the metadata is inconsistent and this code can move funds, treat this as suspicious until you perform the above checks and validate all configuration and endpoints.
功能分析
Type: OpenClaw Skill Name: ai-news-bnb-trader Version: 0.1.0 The skill is a legitimate news-driven trading bot for the Binance Smart Chain (BSC) with a focus on safety and risk management. It includes security-conscious features such as a log redactor to prevent leaking private keys or API tokens (src/utils/logger.ts), an AES-256-GCM encryption utility for local key storage (src/wallet/key-manager.ts), and a dedicated risk manager that enforces daily loss caps and position limits (src/risk/risk-manager.ts). The code logic is transparent, follows the stated purpose in SKILL.md, and contains no evidence of data exfiltration, malicious execution, or prompt injection.
能力评估
Purpose & Capability
The skill claims to be a TypeScript/node BNB trading bot — the source code matches that purpose (news fetcher, signal models, risk gates, DEX integration). However the published metadata / registry entries say 'Required env vars: none' and 'No install spec / instruction-only' while the bundle contains full source and expects sensitive env vars (EVM_PRIVATE_KEY or ENCRYPTED_KEY_PATH + KEY_PASSPHRASE, OPENAI_API_KEY, ONEINCH_API_KEY, RPC URLs, NEWS_API_URL). That metadata omission is an incoherence (sensitive credentials are required for the claimed functionality).
Instruction Scope
SKILL.md and the CLI commands match the code's runtime behavior (start/status/panic/revoke-approvals, key encryption). The instructions request news endpoints, RPCs, and key material which the code consumes. The runtime instructions do not appear to request unrelated system data, but they do direct the agent to perform network calls and on-chain transactions (swap, revoke approvals) when not in dry-run.
Install Mechanism
There is no explicit install spec in the registry (instruction-only), but the package includes package.json and package-lock.json so npm install is required. Dependencies are standard (dotenv, ethers, dev tooling). No remote arbitrary downloads or obscure install URLs are present; risk here is typical for an npm package.
Credentials
The code requires highly sensitive secrets to operate: an EVM private key (or encrypted key + passphrase), RPC URLs (PRIVATE_RPC_URL or RPC_URL), and optional API keys (OPENAI_API_KEY, ONEINCH_API_KEY). Those credentials are proportionate to a trading bot, but the registry metadata does not declare them. Also logger redacts only some envs (EVM_PRIVATE_KEY, OPENAI_API_KEY, ONEINCH_API_KEY) but not KEY_PASSPHRASE or RPC auth tokens — potential leakage points. The skill should have declared required env vars in metadata and documented minimal-permission recommendations.
Persistence & Privilege
The skill does not request always:true, does not modify other skills, and only writes to its own state directory (state-data) and to an encrypted key file if the CLI is used. It will persist its bot-state and optionally an encrypted key file; it will also send transactions that can move funds. Autonomous invocation (default) is present but expected for a skill of this type.
如何使用
  1. 确保已安装 OpenClaw(本地或 Docker 部署)
  2. 在对话框中输入安装命令:/install ai-news-bnb-trader
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /ai-news-bnb-trader 触发
  4. 根据 Skill 的参数说明提供必要输入,即可获得结构化输出
版本历史
v0.1.0
Initial release
元数据
Slug ai-news-bnb-trader
版本 0.1.0
许可证
累计安装 0
当前安装数 0
历史版本数 1
常见问题

AI News BNB Trader 是什么?

TypeScript (Node.js 20+) AI news-driven BNB strategy trading bot for BSC. Use when user wants event/news sentiment signals, strict risk controls, and automat... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 283 次。

如何安装 AI News BNB Trader?

在 OpenClaw 或 Claude Code 对话框中运行命令「/install ai-news-bnb-trader」即可一键安装,无需额外配置。

AI News BNB Trader 是免费的吗?

是的,AI News BNB Trader 完全免费(开源免费),可自由下载、安装和使用。

AI News BNB Trader 支持哪些平台?

AI News BNB Trader 跨平台运行,可在任意部署了 OpenClaw / Claude Code 的环境中使用(cross-platform)。

谁开发了 AI News BNB Trader?

由 happyRstudent(@happyrstudent)开发并维护,当前版本 v0.1.0。

💬 留言讨论