TradeOS enables agents to manage centralized exchange accounts via natural language. Built on CCXT, it supports API key management (encrypted), spot and derivatives orders (market/limit/stop/take-profit), cross-exchange portfolio aggregation, DCA scheduling, arbitrage and funding-rate scanners, conditional orders, anomaly detection and automated security reports. The skill defines safe workflows (confirmation steps, no withdrawals allowed) and operational guardrails.
Use this skill when an agent must interact with exchange accounts: add/list/remove API keys, query balances, place trades with explicit user confirmation, set up DCA plans, monitor arbitrage/funding rates, configure alerts, or generate periodic PnL and security reports. It should be loaded for any user intent mentioning trading, orders, balances, portfolio, DCA, arbitrage, or security.
Best for agents that can make network calls and interact with external APIs (agents with filesystem/network permissions and CCXT-capable backends).
TradeOS is a comprehensive CEX trading and portfolio management skill supporting 100+ exchanges via CCXT. It features AES-256-GCM encrypted API key storage, multi-exchange trading, DCA strategies, arbitrage scanning, conditional orders, anomaly detection, and security reporting. All 15 bundled scripts are TypeScript and could not be executed by the auditor. The SKILL.md is thorough and well-organized. Security practices are solid (encrypted vault, file permissions 0o600, withdraw-permission rejection, auto-desensitization), though handling of financial API keys and private keys (HyperLiquid) inherently carries risk, and DCA/conditional orders auto-execute after initial authorization.
typescriptccxtbetter-sqlite3Well-built trading skill with impressive feature breadth. SKILL.md is written primarily in Chinese which limits accessibility. The security model is thoughtfully designed (encrypted vault, withdraw-permission rejection, risk guard, manual approval hooks), though financial skills always carry elevated risk. No malicious behavior detected. Scripts couldn't be run due to TypeScript format but code review shows clean, well-structured implementation.