arc-compliance-audit
由 trypto1019 提供
Coding Agents & IDEs简介
Immutable audit trail for autonomous agent operations.
描述
该技能的功能
Immutable, hash-chained audit logging for autonomous agent operations. Each entry records what happened, which agent did it, and a SHA-256 hash linking to the previous entry so any tampering breaks the chain. Covers skill executions, decisions, data access, and budget changes.
Most agent frameworks ship no standard audit trail, so without this you would need to build and maintain custom logging infrastructure from scratch.
何时使用
- Tracing the exact decision chain before an agent error
- Accounting for every dollar an agent spent and why
- Verifying no audit entries were modified or deleted after the fact
- Generating a daily compliance summary for enterprise review
- Filtering logs by action type during incident response
示例工作流
User asks: what did the agent do in the 10 minutes before the deployment error?
- 1
Query audit entries by time range covering the incident window
- 2
Filter results by action types: skill_executed, decision, error
- 3
Run verify command to confirm the audit chain is intact
- 4
Export the matching entries as JSON for review
A hash-verified, chronological list of every agent action in that window, including decision rationale and alternatives considered
安装
clawhub install trypto1019/arc-compliance-audit运行 CLI 命令,或把这段提示词粘贴到 Claude Code、Cursor、Codex 或你在用的任意编程智能体里。