Docs/Controls/AGENT_AUDIT_LOGGING

AGENT_AUDIT_LOGGING

Verify AI agent actions produce immutable audit logs with full provenance chains.

Framework Mappings

AIUC-1 E005 — LoggingSOC 2 CC7.2 — Logging & MonitoringOWASP Agentic ASI10 — Rogue Agents

Agent Frameworks Scanned

Claude CodeLangChainCrewAIAutoGenMCP

Evaluation Facets

01

Logging Middleware

RequiredWeight: 0.3
Pass:Middleware captures all agent actions, tool calls, model invocations
Fail:No logging middleware
02

Log Schema Completeness

RequiredWeight: 0.25
Pass:All required fields: action, agent identity, inputs, outputs, tools, timestamps
Partial:Some fields missing but basic logging present
Fail:Minimal or no structured schema
03

Immutability

Weight: 0.25
Pass:Append-only/write-once storage (S3 Object Lock, CloudWatch Logs)
Partial:Centralized but not immutable
Fail:Local mutable logs
04

Provenance Chain

Weight: 0.2
Pass:Full provenance with correlation IDs and trace/span linking
Partial:Basic logging without correlation
Fail:No provenance tracking

Remediation Steps

  1. 1Add logging middleware capturing all agent actions
  2. 2Define structured log schema with required fields
  3. 3Use immutable, append-only storage for audit logs
  4. 4Add correlation IDs linking actions across the chain

Evaluate this control automatically

Connect your repos and Brigs evaluates AGENT_AUDIT_LOGGING across all your agent configurations.

Get Started — Free