• 🧱 NostrTrust Trust Scoring System Draft (Model V0.2)
  • 🔁 Overview of Key Rules
    • ✅ Users with T Score ≥ 1000 Can Earn Points by Posting Content
    • ❌ Penalties for Spam or Misconduct
  • 🧮 Example Scenarios
    • Scenario One: A High-Trust User Posts Normally
    • Scenario Two: A High-Trust User Posts Spam
  • 🛠 Implementation Suggestions
    • 1. Content Classification Tagging System (AI-Assisted)
    • 2. Reporting Process Loop Design
    • 3. Point Change Notification System
  • 📋 Initial Trust Setup
  • ✅ Summary (V0.2 Update Highlights)
• 🧱 NostrTrust 信任积分体系初稿（模型 V0.2）
  • 🔁 新增规则说明
  • 🧮 示例场景
    • 场景一：高分用户正常发帖
    • 场景二：高分用户发垃圾帖
  • 🛠 实现建议
    • 1. 内容分类标签系统（AI 辅助）
    • 2. 举报流程闭环设计
    • 3. 积分变动通知系统
  • 📋 初始信任设定
  • ✅ 小结（V0.2 更新重点）

§🧱 NostrTrust Trust Scoring System Draft (Model V0.2)

✨ Version: V0.2
📅 Date: May 18, 2025
🧭 Objective: To establish a decentralized trust scoring system driven by high-quality interactions for members of the Nostr community, including developers and content creators.

§🔁 Overview of Key Rules

This version introduces a stricter mechanism to ensure that only high-trust users (T Score ≥ 1000) can influence others’ scores through their actions. This helps maintain content quality and prevents abuse from low-reputation accounts.

§✅ Users with T Score ≥ 1000 Can Earn Points by Posting Content

💡 Posts from new users do not contribute to score growth but serve as discovery and interaction opportunities.

§❌ Penalties for Spam or Misconduct

All users, including those with high scores, will face penalties if they engage in spam or misconduct:

⚠️ If a user’s T Score drops below 1000 due to penalties, they will no longer receive points for posting until they recover above 1000 points.

§🧮 Example Scenarios

§Scenario One: A High-Trust User Posts Normally

• User A has a T Score of 2000
• Posts an original piece → +10 points
• Receives a like from a user with a T Score of 1200 → +2 points
• Receives a comment from a user with a T Score of 900 → No points
• Receives another comment from a user with a T Score of 1500 → +3 points
• Gets shared by a user with a T Score of 1000 → +5 points
  ✅ Total points earned this time: +20 points

§Scenario Two: A High-Trust User Posts Spam

• User B has a T Score of 1500
• Posts an obvious advertisement → Gets reported and verified → -30 points
• User C, with a T Score of 10,000, likes this content → Does not add points (because the content is marked as spam)

§🛠 Implementation Suggestions

§1. Content Classification Tagging System (AI-Assisted)

Use lightweight NLP models to classify each piece of content:

• ad (advertisement)
• spam (spam)
• duplicate (duplicate)
• sensitive (sensitive)
• normal (normal)

Tags are automatically analyzed by DVM and recorded as kind=10002 events.

§2. Reporting Process Loop Design

User reports → Recorded as kind=10003 → DVM triggers preliminary assessment → If more than 3 reports → Submitted to DAO quick vote → Approved after review → Deduct points

§3. Point Change Notification System

All point changes should be notified to the user via Nostr Event (such as kind=10001), and displayed in detail on the client side.

§📋 Initial Trust Setup

✅ All initial trusted users’ T Scores are certified by DVM or DAO multisig and recorded as kind=10000 events.

§✅ Summary (V0.2 Update Highlights)

In this version, we introduced the following key mechanisms:

• Only high-trust users (T Score ≥ 1000) can give points via likes, comments, and shares
• The shared content owner must also be a high-trust user to receive share points
• Encourages high-quality content creation
• Prevents low-reputation users from affecting score fairness
• Strengthened detection and penalty mechanisms for spam content

§🧱 NostrTrust 信任积分体系初稿（模型 V0.2）

✨ 版本号：V0.2
📅 初稿日期：2025年5月18日
🧭 目标：构建一个以高质量互动驱动的去中心化信任评分系统，适用于 Nostr 社区成员、开发者与内容创作者。

§🔁 新增规则说明

§✅ 高分用户发帖可获得积分（激励优质内容）

T Score ≥ 1000 的用户可以发布原创内容来获取积分：

💡 注：新用户发布的内容不计入积分增长，仅用于被发现和互动。

§❌ 垃圾内容或不良行为会被扣分（反滥用机制）

所有用户（包括高分用户）若发布以下类型内容，将触发自动或人工审核，并扣除相应积分：

⚠️ 如果用户 T Score 因扣分降至 1000 分以下，其后续发帖不再获得主动积分，直到重新恢复至 1000 分以上。

§🧮 示例场景

§场景一：高分用户正常发帖

• 用户 A，T Score = 2000
• 发布一篇原创内容 → +10 分
• 被 T Score 为 1200 的用户点赞 → +2 分
• 被 T Score 为 900 的用户评论 → 不得分
• 被 T Score 为 1500 的用户评论 → +3 分
• 被 T Score 为 1000 的用户转发 → +5 分
  ✅ 总计本次发帖获得：+20 分

§场景二：高分用户发垃圾帖

• 用户 B，T Score = 1500
• 发布一条明显广告帖 → 被举报并核实 → -30 分
• 用户 C，T Score = 10000 → 点赞该内容 → 不再加分（因内容已被标记为垃圾）

§🛠 实现建议

§1. 内容分类标签系统（AI 辅助）

使用轻量级 NLP 模型对每篇内容进行分类打标签：

• ad（广告）
• spam（垃圾）
• duplicate（重复）
• sensitive（敏感）
• normal（正常）

标签由 DVM 自动分析并记录为 kind=10002 的事件。

§2. 举报流程闭环设计

用户举报 → 记录到 kind=10003 → DVM 触发初步评估 → 若超过 3 次举报 → 提交 DAO 快速投票 → 审核通过后 → 扣除积分

§3. 积分变动通知系统

所有积分变动应通过 Nostr Event 通知用户本人（如 kind=10001），并在客户端展示积分变化详情。

§📋 初始信任设定

✅ 所有初始可信用户的 T Score 由 DVM 或 DAO 多签认证后写入链上事件（kind=10000）。

§✅ 小结（V0.2 更新重点）

在本版本中，我们新增了以下关键机制：

• 只有来自高分用户（T Score ≥ 1000）的点赞、评论和转发才能给接收者增加积分
• 被转发人也必须是高分用户（T Score ≥ 1000）才能获得转发积分
• 继续鼓励优质内容输出
• 防止低质量互动影响积分公平性
• 强化垃圾内容识别与惩罚机制