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
- Before AI-system decision logs or evidence archives must satisfy LGD three-laws compliance — as a front gate check.
- In electronic data forensics, when verifying hash-chain continuity and preventing tampering.
- In audits, when checking whether subject identity and timestamp logic are consistent and complete.
- When a zero-dependency compliance check with no extra third-party libraries is required.
Usage
- Main command:
python evidence_chain_check.py --verify all - Selected checks: use
--mode time(time monotonicity),--mode entity(subject consistency) or--mode hash(hash chain) to trigger each one separately. - Input data: point
--input <path>at the evidence file under check.
Workflow
- Prepare the evidence data source containing timestamps, signatures and hash values.
- Run the Python command to load
evidence_chain_check.pyinto the runtime. - Select or enable the three-laws check mode(s) as needed (time / subject / hash).
- The system compares the input data automatically, detecting monotonicity consistency and chain-break risk.
- Emit a compliance report flagging anomalies and potential evidence-integrity gaps.
- Close the loop: archive, repair or block based on the findings.
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
- AI 系统决策日志或证据归档需满足 LGD 三律合规前，进行前置校验。
- 电子数据取证中验证哈希链连续性、防止篡改时调用。
- 审计场景中检查主体身份与时间戳逻辑是否一致且完整时。
- 需要零依赖且无需额外第三方库的合规校验工具执行环境时。
用法
- 主命令：
python evidence_chain_check.py --verify all - 指定校验项：支持
--mode time（时间单调）、--mode entity（主体一致）、--mode hash（哈希链）分别触发。 - 输入数据：通过
--input <路径>参数指向待检的证据文件。
流程
- 准备包含时间戳、签名及哈希值的证据数据源。
- 执行 Python 命令加载
evidence_chain_check.py脚本进入运行环境。 - 根据需求选择或启用三律校验模式（时间/主体/哈希）。
- 系统自动比对输入数据，检测单调性一致性与链断裂风险。
- 输出合规报告并标记异常项与潜在证据完整性漏洞。
- 依据检测结果执行归档、修复或阻断操作闭环。
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。