okxprediction

Predicts short-term BTC market direction using multi-timeframe analysis, funding rates, and sentiment to signal trade execution, watch, or no trade.

Security Scan

The skill's trading logic is coherent, but it omits where to get market data and requests no OKX/trading credentials despite claiming suitability for OKX — an incoherence that could lead to surprising behavior if combined with other agent capabilities.

Purpose & Capability

The skill is presented as an OKX-focused BTC short-term prediction system (meta and description say '适用于OKX合约'), but the package declares no required environment variables, no API keys, and contains no instructions for connecting to OKX or any market data provider. If the skill is intended to generate signals only (not execute trades), the OKX claim is misleading; if it intends to execute trades, it lacks any credential handling or API-call instructions.

Instruction Scope

SKILL.md contains a detailed multi-timeframe scoring methodology and a fixed JSON output schema — it does not instruct the agent to read local files, access system config, or call external endpoints. However it also does not specify how to obtain the required inputs (K-line data, funding rate, long/short ratio, liquidation zones). That gap means an agent may attempt to fetch data from arbitrary sources or combine this skill with other connectors to obtain data/credentials.

Install Mechanism

Instruction-only skill with no install spec and no included code files. This minimizes on-disk execution risk; there are no downloaded binaries or external installers to review.

Credentials

No environment variables or credentials are declared, yet the meta description explicitly targets OKX contracts. For a skill expected to integrate with an exchange, the absence of declared API keys (OKX_API_KEY, secrets, etc.) is disproportionate and ambiguous. This could either indicate the skill is strictly signalling-only (then the OKX label is misleading) or that it expects the agent to obtain credentials elsewhere (which is risky).

Persistence & Privilege

always is false (normal) and no system-wide config changes are requested. Autonomous invocation is allowed by default — while not problematic on its own, combined with the credential/data sourcing ambiguity it raises the possibility that the agent could autonomously call other connectors or skills to trade using stored credentials. There's no explicit instruction to do so in SKILL.md.

What to consider before installing

What to consider before installing: - The skill's logic and scoring are self-contained and sensible as a signalling methodology, but it does not say where to get the market inputs (candles, funding rate, long/short ratio) or how it would connect to OKX. Decide whether you expect this to be a pure signal generator (manual use) or an automated trader — the skill as packaged does not include any credential handling or exchange API calls. - Do not supply API keys or exchange credentials to this skill unless you fully trust the author and have reviewed explicit code that uses those keys. If you plan to enable automated trading, require the skill to declare which environment variables it will use (e.g., OKX_API_KEY/SECRET) and document exact API endpoints/calls. - Because the source is 'unknown' and there is no homepage or code to audit, prefer using this only for manual, offline signal generation or after thorough review. Test outputs first against historical/paper data before committing real funds and ensure mandatory risk controls (stop-loss, position sizing) are enforced externally. - Recommended improvements for the author: state data sources and authentication requirements in SKILL.md, optionally add a safe 'dry-run' mode, and provide examples showing how to feed data into the skill (or explicitly declare it expects the agent to provide market data).

OKX预测技能

🎯 技能目标

用于预测BTC短线（1M~1H）及1天级别方向，并输出是否执行交易。

核心逻辑： 不是预测价格，而是预测市场行为与概率方向。

---

📊 输入数据

• K线数据：1D / 4H / 1H / 15M / 5M / 1M
• 资金费率（funding rate）
• 多空比（long/short ratio）
• （可选）爆仓数据

---

🧩 多周期分析逻辑

① 方向层（1H / 1D）

偏多：

• 高点抬高 + 低点抬高
• 价格在均线上方

偏空：

• 高点降低 + 低点降低
• 价格在均线下方

中性：

• 均线缠绕
• 区间震荡

---

② 结构层（15M / 4H）

趋势：

• EMA排列清晰
• 持续创新高/低

震荡：

• EMA缠绕
• 区间波动

突破准备：

• 波动收缩
• 靠近关键位

反转：

• 结构被破坏

---

③ 位置层（5M）

做多：

• 在支撑上方
• 回踩不破

做空：

• 在压力下方
• 反弹不过

禁止：

• 区间中部

---

④ 情绪层

资金费率：

• 0.01 → 多头过热 → 偏空

• < -0.01 → 空头过热 → 偏多

多空比：

• 多头 > 70% → 看跌
• 空头 > 70% → 看涨

---

⑤ 流动性层

• 价格倾向移动到爆仓密集区
• 作为目标位参考

---

⑥ 触发层（1M）

做多：

• 更高低点形成
• 突破并站稳

做空：

• 更低高点形成
• 跌破并站稳

---

🧮 评分系统

项目	分数
方向	20
结构	20
位置	20
情绪	15
流动性	15
触发	10

---

✅ 执行规则

• ≥80 → 执行交易
• 60~79 → 观察
• <60 → 不交易

---

⚠️ 强制规则

• 方向不一致 → 不交易
• 没有位置 → 不交易
• 没有触发 → 不入场
• 必须有止损

---

🔁 执行流程

1. 判断1H/1D方向
2. 判断15M/4H结构
3. 判断5M位置
4. 判断情绪（资金费率+多空比）
5. 判断流动性目标
6. 判断1M触发
7. 计算评分
8. 输出信号

---

📤 输出格式

{
  "direction": "long / short / neutral",
  "score": 0,
  "signal": "execute / watch / no_trade",
  "entry_type": "pullback / breakout / reversal",
  "stop_loss": "price",
  "target": "price zone",
  "reason": "summary"
}