solsticestudioai/arc-t-defi-pack

ARC-T DeFi决策遥测包（样本）是一个合成的DeFi风险到执行决策遥测数据集，用于自主交易研究、链上异常检测和强化学习环境。每条记录代表一个完整的风险到执行生命周期，包括触发市场异常、市场状态快照、对抗性规划推理（MCTS分支计数和获胜策略）、逐步执行链（每步的gas、滑点、延迟遥测）、执行摘要以及最终决策结果和PnL变化。该数据集由SolsticeAI构建，是一个更大商业包的免费样本，100%合成，不包含真实交易、钱包、地址或链上身份，适用于模型训练、竞争评估和公共基准工作。数据集包含10,000个生命周期，覆盖多种决策结果、链和协议，并包括多种风险触发因素。

The ARC-T DeFi Decision Telemetry Pack (Sample) is a synthetic DeFi risk-to-execution decision-telemetry dataset for autonomous-trading research, on-chain anomaly detection, and reinforcement-learning environments. Each row represents a complete risk-to-execution lifecycle: a triggering market anomaly, market-state snapshot, adversarial-planner reasoning (MCTS branch count + winning strategy), step-by-step execution chain with per-step gas / slippage / latency telemetry, execution summary, and final decision outcome with PnL delta. Built by SolsticeAI as a free sample of a larger commercial pack, it is 100% synthetic with no real trades, wallets, addresses, or on-chain identities, making it safe for model training, competitive evaluation, and public benchmark work. The dataset includes 10,000 lifecycles, stratified across decision outcomes, chains, and protocols, and covers various risk triggers.