Soccer Predict

Football match betting prediction system. Auto-scrapes data from titan007.com (Asian handicap, over/under, European odds, fundamentals, lineups, corners, hal...

Security Scan

The skill's stated purpose (scraping titan007 and producing betting predictions) mostly matches its instructions, but the runtime instructions ask the agent to execute JavaScript in browsed pages and persist framework/history files under the user's home directory without declaring those config paths — this mismatch and the use of in-page JS extraction are worth caution.

Purpose & Capability

Name/description (football betting predictions from titan007) align with the instructions: the SKILL.md explains scraping titan007, running a 5-step prediction framework, and producing outputs. No unrelated credentials or external services are requested in metadata.

Instruction Scope

Instructions tell the agent to use browser-based scraping (browser navigate / browser act / browser console exec) and to run JavaScript in the page context to extract data. That is reasonable for scraping, but executing arbitrary JS in a browsed page may access page-local secrets (cookies, localStorage) and could exfiltrate data. The instructions also direct saving model/framework/data across sessions (see persistence), which broadens the skill's runtime footprint beyond a one-off scrape.

Install Mechanism

No install spec and no code files — instruction-only. This is the lowest-risk install model and is consistent with an agent-driven scraping/prediction skill.

Credentials

The skill declares no required environment variables or config paths, which fits its purpose. However, SKILL.md and references explicitly instruct writing persistent files to ~/.claude/projects/*/memory/..., a config path not declared in metadata. The skill requests no secrets, but the combination of in-browser JS and implicit file writes is notable.

Persistence & Privilege

The skill directs the agent to 'save the updated framework and match history to memory files' under ~/.claude/projects/*/memory/... to persist learning across sessions. The metadata did not declare required config paths or warn about persistent file writes. Persisting scraped data and learned parameters to the user's home directory increases long‑term risk (sensitive data retained, larger attack surface) and should be made explicit to users before installation.

What to consider before installing

This skill is coherent with its description, but exercise caution before enabling it. Key points to consider: - The skill instructs the agent to open titan007 pages and run JavaScript in the page context to extract tables. That can access page cookies/localStorage or other site-local data — only run this if you trust the skill source and understand the site’s data sensitivity and terms of service. - The SKILL.md tells the agent to persist learned models and match history to files under ~/.claude/projects/*/memory/..., but the registry metadata does not declare these config paths. Confirm where data will be written, review those files, and ensure you are comfortable with persistent storage of scraped data. - Scraping titan007 may violate that site’s terms or local laws; check legal/ToS implications and betting regulations in your jurisdiction. - Because the skill can store and reuse historical predictions, review stored content for sensitive or personally identifying data and consider running the skill in a sandboxed environment (isolated user account or container) if you want to limit exposure. - Avoid providing any unrelated credentials; this skill does not declare needed secrets. If additional env vars or tokens are requested later, treat that as a red flag. If you decide to proceed: start with a short test in concise mode, monitor network activity and created files, and inspect saved memory files before allowing continued autonomous use.

足球博彩预测 / Football Betting Prediction

三大工作流：数据采集 → 预测分析 → 赛后复盘

输出模式选择

用户请求预测时，询问或推断偏好的输出格式：

• 简洁模式 (Concise): 快速结果 - 最佳推荐、概率、EV、预测比分
• 可视化模式 (Visual/Detailed): 完整 HTML 报告，含数据表格、公式、图表

未指定时：默认使用可视化模式以获得最佳用户体验。

---

工作流一：赛前数据采集

用户提供比赛 ID 或比赛描述时，从 titan007.com 采集数据。

1. 构建 URL: https://zq.titan007.com/analysis/{match_id}cn.htm
2. 使用 browser navigate 打开页面
3. 使用新球体育数据提取数据。详见 references/data-collection.md

关于首发阵容：首发阵容通常在开赛前30-60分钟公布。提前收集时标注"阵容待公布"，使用现有阵容深度信息。接近开球时如有需要可重新检查。

需采集数据点：

• 比赛信息（球队、联赛、时间、场地、天气）
• 亚盘（让球盘）："即"和"早"行，所有机构
• 大小球盘："即"和"早"行，所有机构
• 欧赔（胜平负）："即"和"早"行，前10+机构
• 球队基本面（近期战绩、主客场、历史交锋、联赛排名）
• 大小球增强数据：半场进球、角球
• 首发阵容（如比赛前30-60分钟内可用）

采集完成后，立即进入工作流二。

---

工作流二：预测分析

对收集的数据运行五步预测框架。详见 references/prediction-framework.md

输出模式决定呈现方式：

• 简洁：基于文本的快速摘要，关键数字
• 可视化：完整 HTML 报告，数据表格、概率条、公式展示

步骤概述：

1. 数据整理 - 将所有收集的数据分类整理
2. 基本面分析 - 盘口合理性、走势追踪、机构意图、欧亚转换
3. 盘口概率计算 - 计算亚盘和大小球的真实隐含概率
4. 模型预测 - 两个逻辑回归模型：亚盘模型（权重：盘口0.35 > 基本面0.20+0.20 > 阵容0.15 > 战意0.10）和大小球模型（含xG、联赛因子、半场进球、角球等）
5. EV 计算与推荐 - 计算每个投注选项的期望值，输出最佳推荐和预测比分

默认初始权重：基于 AI 概率判断设置，通过赛后复盘自动优化。

输出格式：

• 亚盘分析及其赢盘概率
• 大小球分析及其赢盘概率
• 所有选项的 EV 值
• 最佳投注推荐
• 预测比分

---

工作流三：赛后复盘

用户提供比赛结果或请求复盘时触发。详见 references/review-framework.md

流程：

1. 回顾本场比赛的所有赛前数据和预测分析
2. 运行偏差分析，比较预测与实际结果
3. 识别预测错误的根本原因
4. 自动优化：根据预测误差调整特征权重
5. 将更新后的框架保存到内存文件以持久化
6. 输出所有已分析比赛的累计准确率统计

持久化：将框架更新和比赛历史保存到内存文件，以便跨会话学习。

目标：亚盘和大小球预测准确率均达到 70% 以上。

---

参考文档

• data-collection.md - 数据采集指南
• prediction-framework.md - 预测分析框架
• review-framework.md - 赛后复盘与学习框架