Audit Facts

Updated at:

Audit Facts is the raw behavior browsing and analysis entry point of the Audit module. It includes three tabs: Session Browsing, Token Analysis, and Behavior Analysis. This topic describes the features and usage of the Audit Facts page.

Page overview

Audit Facts directly describes raw sessions, model requests/responses, tool calls, tool results, and runtime behaviors reported by the Agent. It does not depend on the risk audit switch being enabled. When the risk audit switch is off, Risk Audit and Entity Investigation may only display historical risk data, but Audit Facts remains available for verifying ingestion results, troubleshooting context, and analyzing Token/behavior trends.

The top of the page provides unified filters: data scope, Agent type, and query time range. The data scope currently provides AI Application Audit and AI Runtime Audit. The default query time range is the last 7 days.

Page structure

The Audit Facts page contains three tabs:

Tab

Description

Session Browsing

Filter sessions by host, user, Agent application, source IP, Tool, Session ID, and more. View context timelines.

Token Analysis

Analyze input tokens, output tokens, total tokens, model distribution, time trends, and high-consumption sessions.

Behavior Analysis

View Agent activity, model requests, tool calls, command execution, external interactions, errors, and Session rankings from a neutral behavior perspective.

Session Browsing

Session Browsing uses a left-right split layout:

  • Left panel: Session list showing sessions that match filter conditions.

  • Right panel: Full context details of the selected Session.

Top-level statistics

The top of the page displays overall statistics for the current filter scope:

Metric

Description

Sessions

Total number of matching sessions

Events

Total raw event count

Turns

Number of user-Agent interaction rounds

Steps

Number of Agent execution steps

LLM Requests

Number of LLM model requests

Input Tokens

Total model input token count

Output Tokens

Total model output token count

LLM Responses

Number of LLM model responses

Tool Calls / Results

Number of tool calls and return results

Filter conditions

Session Browsing supports the following filter dimensions:

Filter

Description

Time range

Located in the top-level unified filter. Default is the last 7 days.

Agent type

Located in the top-level unified filter.

Host

Filter by Agent runtime host.

User

Filter by user dimension.

Agent application

Filter by application name.

Source IP

Filter by collection source IP.

Tool

Filter by tools used.

Session ID

Filter or locate a session by specific Session ID.

Filter inputs support prefix queries and return matching values with session counts.

Session list

The Session list displays summary information for matching sessions:

Field

Description

Session ID

Unique session identifier

Tool summary

Main tools used in the session

Host / Agent application

Runtime host and application name

User / Agent type

User identifier and Agent type

Source IP

Collection source IP

Events

Total raw event count

Turns

Interaction round count

Steps

Execution step count

Req / Resp

LLM request count / response count

Tool Calls / Results

Tool call count / result count

Last active

Timestamp of the most recent event

The list is paginated. Click a row to display the Session context in the right panel.

Session context details

After selecting a Session, the right panel displays the complete event timeline for that Session.

Event types

The timeline displays the following raw event types:

Event type

Label

Description

Other event

other

Session-level metadata events

LLM Request

Request

Request sent to the model, including messages input

LLM Response

Response

Response returned by the model, including assistant output

Tool Call

tool call

Tool invocation initiated by the Agent, including tool name and parameters

Tool Result

tool result

Return result from tool execution

Event information

Each event displays the following information:

  • Event ID: Globally unique identifier.

  • Turn / Step: The interaction round and step number the event belongs to.

  • Timestamp: The precise time the event occurred.

  • Content summary: Key content displayed based on event type (user input, model output, tool parameters and results, and more).

Token Analysis

Token Analysis observes model call costs and context consumption, answering: "Which models, applications, or sessions consumed the most tokens, and is consumption concentrated in a few long sessions or anomalous requests?"

Token Analysis typically includes:

Module

Description

Token overview

Displays input tokens, output tokens, and total tokens.

Model distribution and trends

Displays token consumption by model and time trends.

Session ranking

Displays high-consumption sessions with navigation to Session Browsing for context review.

Token bucket distribution

Observes long context and anomalous consumption by Turn or Session token consumption ranges.

Token Analysis is an audit facts view and does not represent a confirmed security risk. If anomalous consumption is detected, combine Session Browsing and Behavior Analysis to confirm whether loop calls, repeated tool failures, excessively long context, or anomalous model requests exist.

Behavior Analysis

Behavior Analysis observes Agent activity from a neutral behavioral perspective. It does not directly equate to risk detection results. It answers: "Which applications are most active, which tools are frequently called, are there large volumes of command execution, external interactions, or tool errors?"

Behavior Analysis typically includes:

Module

Description

Activity metrics

Displays active Agents, session count, model requests, total behavior volume, and other overviews.

Tools and commands

Displays tool call, command execution, command families, and tool error statistics.

External interactions

Displays external target, domain, or network interaction related behaviors.

Session ranking and timeline

Locates sessions with anomalous behavior volumes and navigates to Session Browsing for context replay.

Behavior Analysis helps users understand normal behavior baselines and can provide background for risk investigations. For example, when Risk Audit detects sensitive data exfiltration, Behavior Analysis can show whether the same application or host simultaneously exhibits anomalous external interactions or tool errors.

Cross-page navigation and context positioning

Audit Facts serves as the context viewing entry point for other audit pages. When you need to view raw sessions from Risk Audit, Entity Investigation, or Audit Management, navigate to Audit Facts and view the corresponding Session event timeline in Session Browsing.

When navigating from Risk Audit, use the Session information in risk details to return to the raw session and view model requests, model responses, tool calls, and tool results around the time of the risk. When navigating from Entity Investigation, view the session context where a specific entity appeared. When navigating from Audit Management, verify whether a specific application or host is producing audit data.

This capability supports investigation path continuity: you can navigate from risks, entities, or ingestion status back to raw audit facts. The Audit Facts page itself only provides browsing, filtering, and context viewing capabilities, without risk annotation or disposition functions.

Use case examples

Example 1: View raw context from a security event

  1. View the details of a risk in Risk Audit.

  2. Note or navigate to the associated Session of the risk.

  3. In Session Browsing under Audit Facts, view the raw event timeline for that Session.

  4. Combine model requests, model responses, tool calls, and tool results to understand the complete process around the security event.

Example 2: Investigate anomalous token consumption

  1. In Token Analysis under Audit Facts, identify a Session with anomalously high consumption.

  2. Navigate to Session Browsing, or use the Session ID filter to locate the Session.

  3. Review LLM Request/Response events to check for excessively long context or repeated calls.

  4. Combine tool call events to confirm whether loop calls or failure retries exist.

Example 3: Verify ingestion data completeness

  1. On the Audit Facts page, filter by recent time range and target Agent type.

  2. Review the top-level statistics in Session Browsing to confirm that Session count and Event count match expectations.

  3. Open a specific Session to view the event timeline and confirm that Turns, Steps, Request/Response, and Tool Call/Result events are complete.

  4. If events are missing or fields are incomplete, check the ingestion configuration and collection component status.