AI Agent Audit Report

Executive Summary

• Overall risk:
• Audit mode:
• Highest-risk failure mode:
• Fastest useful fix:
• Confidence:

System Map

Describe the agent workflow: inputs, context assembly, model calls, tools, memory, side effects, evals, and human handoffs.

Evidence Reviewed

-

Findings

1. Finding Title

Severity:

Laws:

Evidence:

Why it matters:

Fix:

Verification:

Owner:

Confidence:

What Looks Solid

-

Unknowns / Needed Evidence

-

Black-Box Limits

Use this section only when internals were unavailable. State what could not be inspected and which findings are hypotheses.

7-Day Fix Plan

1. 2. 3.

30-Day Hardening Plan

1. 2. 3.