agent-audit-logging

What problem does it solve?

This Skill eliminates the uncertainty and risk of operating multi-agent systems without complete evidence of what actions were taken, why decisions were made, and whether governance rules were followed.

Core Features & Use Cases

• Structured audit event schema: Defines consistent, queryable fields for invocation, tool calls, decisions, responses, errors, handoffs, and token usage.
• Traceability chains for forensics: Builds a trace timeline and parent-child graph to replay agent behavior across a full execution path.
• Compliance reporting and monitoring: Generates compliance summaries (including error and tool-usage analysis) and supports real-time dashboards for operational awareness.

Use case: when a production multi-agent workflow escalates to humans after tool failures, you can export a trace for incident forensics and produce a compliance report showing which rules were met and where failures occurred.

Quick Start

Ask your agent system to enable audit logging for each invocation, tool call, decision, and handoff, then generate a compliance report for a selected time window.