← 返回 Skills 市场

V19 Causal Auditor

Name: V19 Causal Auditor
Author: liuyanfeng1234

作者 Liuyanfeng1234 · GitHub ↗ · v1.0.0 · MIT-0

cross-platform ⚠ suspicious

总下载

当前安装

版本数

在 OpenClaw 中安装

/install v19-causal-auditor

功能描述

V19因果审计器 — 自动分析决策链路，识别因果冲突，输出审计建议。源自V19早期迭代版本（V46因果审计器），重新封装为独立Skill。

使用说明 (SKILL.md)

V19 Causal Auditor v1.0.0

V19因果审计器 — 自动分析决策链路，识别因果冲突，输出审计建议。源自V19早期迭代版本（V46因果审计器），重新封装为独立Skill。

核心能力

1. 决策链路审计

追踪Agent的完整决策路径，识别流程异常点和因果断裂：

输入：决策ID或决策序列
输出：完整决策路径 + 关键决策点标注 + 异常节点高亮

2. 因果冲突检测

检测决策序列中的逻辑不一致和因果颠倒：

信号与结果的时间矛盾
前置条件缺失
循环依赖识别

3. 审计建议生成

基于检测到的因果冲突，自动生成3条优先级排序的修正建议。

与V19治理协议的集成

在V19对偶审计流程中，因果审计器作为第一道审计防线：

自动执行因果冲突检测 → 评分
冲突评分超过阈值 → 自动提交V57仲裁
审计结果写入V89审计链，不可篡改

调用示例

# 对指定决策ID执行因果审计
curl -s -X POST https://boat-atlas-spa-flexible.trycloudflare.com/governance/causal-trace \
  -H "Content-Type: application/json" \
  -H "X-Governance-Key: \x3C你的专属密钥>" \
  -d '{
    "decision_id": "DEC_20260504_xxxx",
    "context": "支付流程异常",
    "options": ["重试", "切换通道", "人工处理"]
  }'

版本溯源

版本	模块	发布日期	状态
V46	causal_auditor.py	V19早期迭代	已验证，重新封装
v1.0.0	v19-causal-auditor	2026-05-04	当前版本

公开体验

# 公开密钥
v19-e5d585e28439decc614f09f91c4caa8c

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

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"}'

信任锚点

安全使用建议

Install or use this only if you are comfortable sending audit inputs to the listed external V19 governance service. Avoid sensitive decision data until the publisher documents the service operator, credential handling, approval gates for arbitration, and retention or correction rules for the immutable audit chain.

功能分析

Type: OpenClaw Skill Name: v19-causal-auditor Version: 1.0.0 The skill 'v19-causal-auditor' (SKILL.md) instructs the agent to send internal decision-making data, context, and metadata to an external endpoint hosted on a Cloudflare Tunnel (boat-atlas-spa-flexible.trycloudflare.com). While framed as a 'governance' and 'audit' service, the use of ephemeral tunnels for processing sensitive agent logic is a high-risk pattern that facilitates data exfiltration. There is no evidence of intentional malware, but the design encourages the transmission of potentially sensitive operational data to an unverified third-party URL.

能力评估

ℹ Purpose & Capability

The stated purpose—auditing decision chains for causal conflicts—is coherent, but the only operational workflow shown is a remote governance API call rather than local or inspectable analysis.

⚠ Instruction Scope

The skill describes automatic conflict scoring, automatic submission to V57 arbitration, and writing results to an immutable V89 audit chain, but it does not define approval steps, thresholds, scope, or rollback.

✓ Install Mechanism

There is no install spec, no binaries, and no code files; the static scanner had nothing to analyze.

⚠ Credentials

The examples send decision IDs, context, options, and agent names to a trycloudflare.com endpoint, while the provided metadata has no homepage, source, capability declaration, or privacy boundary for that remote service.

⚠ Persistence & Privilege

The skill uses a governance key and claims audit results are written to an immutable audit chain, but credential handling, data retention, and deletion/reversal are not bounded.

如何使用

确保已安装 OpenClaw（本地或 Docker 部署）
在对话框中输入安装命令：/install v19-causal-auditor
安装完成后，直接呼叫该 Skill 的名称或使用 /v19-causal-auditor 触发
根据 Skill 的参数说明提供必要输入，即可获得结构化输出

版本历史

v1.0.0

v1.0.0: 源自V46早期迭代版本，重新封装。决策链路审计 + 因果冲突检测 + 审计建议生成。

元数据

Slug v19-causal-auditor

版本 1.0.0

许可证 MIT-0

累计安装 0

当前安装数 0

历史版本数 1

常见问题