agent-trace-audit
About this agent
当用户说『AI刚才干了什么我要查』『agent操作有没有越界』『出事了能回溯吗』『给AI行为留个账本』，或要给 agent 的操作做可审计留痕时使用。把 agent 的动作日志（时间戳/执行者/动作/对象/是否过闸）建成可追溯账本，标出未过闸的越界操作（有籍），输出时间线+违规清单。可运行脚本（trace_audit 审计器）。理论根基：LGD 三律之有籍（全程可追溯）。与 agent-redteam-kit/有门禁互补。触发词：行为留痕、trace audit、操作审计、agent回溯、可追溯、行为账本、审计日志。
Security
This agent was reviewed and published through ClawHub's own moderation and security scanning pipeline before being listed on their registry. ClawHub checks for malicious code, dependency vulnerabilities, and policy violations.
When you run npx @agentpowers/cli install agent-trace-audit, AgentPowers downloads the package to a sandbox and runs a full security scan (ClamAV, Snyk, GuardDog, VirusTotal) before allowing installation. If the scan fails, installation is blocked.
Powered by ClawHub
This agent is sourced from ClawHub, a large open registry of community-built AI tools. By integrating ClawHub into AgentPowers, you get access to hundreds of additional agents while still benefiting from our security scanning, one-command install, and unified management through the AgentPowers CLI.
Getting started
New to AgentPowers? Learn how to install and use agents from the community.