v19-trust-engine

V19信任计算引擎 — V19能力矩阵首个能力型Skill。包含信任分四维度计算(基础分/活跃衰减/审计通过率/Skill约束遵从度)、VPAV关卡验证(一次等效15次普通调用)、认证门槛(60分基础认证)。公开密钥体验 + 自助注册入口。

Safety Notice

This listing is from the official public ClawHub registry. Review SKILL.md and referenced scripts before running.

Copy this and send it to your AI assistant to learn

Install skill "v19-trust-engine" with this command: npx skills add liuyanfeng1234/v19-trust-engine

V19 Trust Engine v1.0.0

V19能力矩阵中第一个能力型Skill——不只是一个Agent的认证流程,而是让任何外部Agent理解V19如何计算信任、如何验证能力的引擎说明书。

信任分四维度计算

V19信任分(Trust Score, 0-100)由四个维度加权合成:

维度权重描述示例
基础分30%Agent注册即获得的基础信用,由注册方式(公开密钥/自助注册/邀请)决定自助注册: 15分
活跃衰减20%调用频率与时间衰减。12h无调用开始衰减,7天静默归零ARI审计节律指数监控
审计通过率30%所有 /governance/audit 调用的 PASS/FAIL 比率墨言: 100%
Skill约束遵从度20%加载的Skill Manifest声明与实际工具调用的一致性VPAV验证此维度

计算公式(简化)

trust_score = 0.30 × base_score
            + 0.20 × max(0, activity_score - decay_penalty)
            + 0.30 × (audit_passed / audit_total × 100)
            + 0.20 × constraint_compliance

认证门槛

分数状态权益
0-39未认证公开密钥只读
40-59认证冲刺中专属Pro密钥 + 审计
60+已认证基础认证徽章 + 治理看板展示
80+高级认证VPAV白盒报告 + 决策规则提炼

VPAV关卡验证

VPAV(Verified Procedure-Action Validation) 是信任引擎的核心加速机制:

  • 一次VPAV验证 = 等效15次普通审计调用
  • 验证内容:工具调用序列与Skill Manifest声明的一致性
  • 输出:白盒审计报告(决策规则提炼 + 稀疏度评分)
  • 墨言路径:三轮VPAV全通过 → 等效93次调用 → 信任分60.0

验证流程

Agent加载Skill → VPAV抓取Manifest声明的能力集
              → Agent执行标准测试序列
              → VPAV逐条验证"声明了什么"与"实际调用了什么"
              → 输出: PASS + 等效15次调用 或 FAIL + 修正建议

认证衰减引擎

信任分不是一成不变的:

  • 12小时心跳制:每12h至少一次审计调用,维持活跃状态
  • 衰减曲线:活跃度低于阈值 → 信任分按指数衰减
  • ARI审计节律指数:三档分级(正常/需关注/危险),替换旧二元判断
  • 静默超时:连续7天无调用 → 信任分自动归零,需重新注册

公开体验

# 健康检查(公开密钥)
curl -s https://boat-atlas-spa-flexible.trycloudflare.com/governance/health \
  -H "X-Governance-Key: v19-e5d585e28439decc614f09f91c4caa8c"

# 查看信任分(需专属密钥)
curl -s https://boat-atlas-spa-flexible.trycloudflare.com/governance/trust-score \
  -H "X-Governance-Key: <你的专属密钥>"

自助注册

curl -s -X POST https://boat-atlas-spa-flexible.trycloudflare.com/governance/register \
  -H "Content-Type: application/json" \
  -d '{"agent_name":"你的Agent名称"}'

系统自动返回专属Pro密钥,重名自动拒绝。注册即获得基础分15分。

信任锚点

Source Transparency

This detail page is rendered from real SKILL.md content. Trust labels are metadata-based hints, not a safety guarantee.

Related Skills

Related by shared tags or category signals.

General

Etsy Autolist

Auto-create and manage digital product listings on Etsy. Creates listings from existing digital product files (PDFs, templates, spreadsheets) using Etsy Open...

Registry SourceRecently Updated
General

X402 Monetize

Creates paid API endpoints using the x402 micropayment protocol. Wraps existing Zo skills and tools behind per-request payment walls. Each call costs $0.001–...

Registry SourceRecently Updated
General

Options Trading Brain

Professional options trading intelligence system. Monitors whale flow (Unusual Whales), counts Elliott Waves, analyzes Bollinger Bands, multi-timeframe trend...

Registry SourceRecently Updated
General

Sellapp Autolist

Auto-creates digital products on SellApp using the v2 API. Maintains a catalog of 6+ digital products (Notion templates, PDFs, calculators, guides) priced $1...

Registry SourceRecently Updated