lgd-crypto-guard

该技能以 Python 命令行工具或模块形式，对 AI 模型、数据源与访问请求进行 LGD“有籍·有证·有门禁”自检，输出结构化报告和 PASS/FAIL 退出码；适用于隐私计算、联邦学习、密文传输与数据最小化场景的部署前合规审查。材料未提供独立安全扫描或评测证据。

zhaoxinghua09-cell

热度 61.72

最近更新 2026-10-17 07:41:38

分类：安全依赖：代码执行应用场景：报告撰写 / 工作流 / 法律文档

安装技能

复制下面的话术发给你的 Agent，它会通过 secure-skills 检索并安装该技能。

执行命令安装此 skill： npx secure-skills-cli add lgd-crypto-guard

Metadata

name: lgd-crypto-guard
description: lgd-crypto-guard — 把 LGD 三律（有籍·有证·有门禁）翻译进隐私计算 / 数据加密与最小化 AI 应用（联邦·密文·脱敏）的合规守门器，对 AI 系统做三律自评与门禁判定（零依赖）。适用场景：当需要隐私计算守门能力时使用本技能；触发词：lgd-crypto-guard。
slug: lgd-crypto-guard
version: 1.0.1
display_name: 隐私计算守门
display_name_en: Privacy-Compute Guard
agent_created: true
author: XLGD · 诺声(Logos)@SynomosAI
license: MIT
license_scope: 代码（.py 文件）MIT；本 SKILL.md 与其中理论文本不在 MIT 覆盖范围内（见文末「License & attribution」）
allowed-tools: Read, Glob, Grep, Bash, Write, WebFetch
category: AI 治理
platforms: ["windows","macos","linux"]
copyright: XLGD · SynomosAI
description_zh: lgd-crypto-guard — 把 LGD 三律（有籍·有证·有门禁）翻译进隐私计算 / 数据加密与最小化 AI 应用（联邦·密文·脱敏）的合规守门器，对 AI 系统做三律自评与门禁判定（零依赖）。
description_en: Translate the LGD three laws into privacy-computing and minimized AI applications (federated, ciphertext, desensitized): three-law self-assessment and gate decisions (zero-dependency).
classification: {"skillhub":["ai-governance"],"iso_25010":["Security","Maintainability"]}
risk_tier: medium
risk_rationale: 门禁批量补丁默认 medium（按默认策略设计为中等风险）：调用 API/读写文件/本地执行，最小权限，密钥走 Secret；若实际涉对外代执行(push/发布/发送/上架)须人工复评，改 high 并走人工批准。
tags: ["lgd-crypto-guard","AI 治理","隐私计算","数据最小化"]
read_when: ["AI 模型部署前要验证数据流转是否符合 LGD 三律","系统处理敏感数据，涉及隐私计算/联邦学习/密文传输且要最小化原则","AI 接口有日志泄露风险或权限分级错误，要自动脱敏与门禁拦截","用户说「这个隐私计算方案合规吗」「明文有没有出安全区」"]
ai_protocol: true
spec: usk/1.0
interface: {"type":"cli","entry_point":"scripts/crypto_guard.py","runtime":"python3","call_pattern":"args"}
input_schema: {"args":{"type":"string","description":"命令行参数透传给主脚本，具体取值见各包 --help","required":false}}
output_schema: {"stdout":{"type":"string","description":"标准输出为 JSON 结果，含逐项检查与总体结论"},"exit_code":{"type":"integer","description":"0=PASS/完成；1=FAIL/有警告或待补强；2=用法或参数错误"}}
capabilities: ["cli_tool"]
permissions: {"network":false,"filesystem":false,"subprocess":false}
requirements: {"python_packages":[],"min_python":"3.8"}
changelog: 1.0.1: Bilingual rewrite (full EN+ZH, English first), added read_when; 1.0.0: initial upload (USK v3)

凡自治之物 · LGD 治理学 · 创造术家族｜评测协议：见平台质量门禁规范

English

When to use

Call it before an AI model deployment, to verify that data flows meet the LGD "registered · evidenced · gated" standard.
Enable it when a system handles sensitive data in privacy-computing, federated-learning or ciphertext-transmission scenarios and must satisfy the data-minimization principle.
Use it when an AI interface shows log-leakage risk or flawed permission tiers, triggering automatic desensitization and gate blocking.

Usage

In a Python runtime, import the core classes and functions from the crypto_guard.py module. Configure the inputs as the data source under check, the model path and the compliance policy level (e.g. HIGH). Recommended invocation:

from crypto_guard import GuardianEngine, LGD_POLICY
guard = GuardianEngine(policy=LGD_POLICY.STRICT)
result = guard.check_model(ai_model_path="local/model.pkl")

Workflow

Environment prep: load the crypto_guard module; confirm no dependency conflicts.
Data assembly: organize the AI model file, training dataset and access-request context into a standardized input object.
Compliance scan: run the evaluation engine for an LGD three-law self-check on data lineage and encryption strategy.
Report generation: emit a structured report with the "registered / evidenced / gated" status verdicts.
Execution control: if the check fails, automatically issue encryption, desensitization or access-denial instructions; on pass, open the service.

Evidence & safety statement

Evidence: TBD (no evaluation / badge link yet; update here once it is produced)
Version: 1.0.1
Security: no independent security scan has been performed (TBD); the conclusion is pending and no endorsement is given until it is completed.

中文

何时用

在 AI 模型部署前需验证数据流转是否符合 LGD“有籍·有证·有门禁”标准时调用。当系统处理敏感数据，涉及隐私计算、联邦学习或密文传输场景且需满足最小化原则时启用。若检测到 AI 接口存在日志泄露风险或权限分级错误，需触发自动脱敏与门禁拦截机制时使用。

用法

在 Python 执行环境中，直接导入 crypto_guard.py 模块中的核心类与函数。需配置输入参数为待检数据源、模型路径及合规策略等级（如 HIGH）。实际调用推荐通过以下结构初始化并发送请求：

from crypto_guard import GuardianEngine, LGD_POLICY
guard = GuardianEngine(policy=LGD_POLICY.STRICT)
result = guard.check_model(ai_model_path="local/model.pkl")

流程

环境准备：加载 crypto_guard 模块，确认无依赖冲突。
数据构建：组织 AI 模型文件、训练数据集及访问请求上下文为标准化输入对象。
合规扫描：调用评估引擎对数据血缘与加密策略进行 LGD 三律自检。
报告生成：输出包含“有籍/有证/门禁”状态判定结果的结构化报告。
执行控制：若校验未通过，自动触发数据加密、脱敏或访问拒绝指令；通过后开放服务。

证据与安全声明

证据：TBD（暂无评测/徽章链接；产生后在此更新实际链接）
版本：1.0.1
安全：未做独立安全扫描（TBD），结论待补；补全前不作推荐背书。