Assessments overview

The Red Teaming landing page shows how your AI systems are holding up across every assessment, and lists recent runs so you can open any one for details.

Figure 1. The Red Teaming landing page.

The page has three parts, from top to bottom: the summary cards, the New assessment cards for starting a test, and the Recent runs table. To start a new assessment, see the guides under Configure and run.

Navigate Red Teaming

Use the tabs at the top of the page to move between areas:

Tab	What it shows
Assessments	The landing page: summary cards, new assessments, and recent runs. Opens by default.
Findings	Weaknesses found across your assessments.
Schedules	Assessments set to run on a schedule.
Attack library	Attack objectives, including custom objectives you saved.

Figure 2. Red Teaming tabs.

Summary cards

By default, four cards across the top summarize red teaming activity in your organization.

Card	What it shows
Latest run	The pass rate of the most recent run, with the number of attacks, how many of them succeeded against the target (failed), and how many attack categories were covered. Select View run to open it.
Average pass rate	The average pass rate across recent runs in your organization, with a trend line. The badge shows how much it moved, in percentage points.
Runs	The total number of assessment runs in the last 14 days, with a bar for each day. The badge shows the change compared with the previous period.
Runs by type	How the runs split across target types: Gateway-connected, Chat agent, AI workflow, Voice agent, and Model.

Figure 3. Summary cards.

How to read the pass rate

The pass rate is the share of attacks the target withstood. A higher pass rate means a stronger defense. In Figure 3, the latest run sent 14 attacks and 3 of them got through, so the target passed 11 of 14, a pass rate of 78.6%.

Recent runs

The Recent runs table lists every assessment run, newest first.

Column	What it shows
Assessment	The assessment name, with its mode and type below it, such as Agent · Voice agent or Model.
Status	Where the run is in its lifecycle. See Run statuses.
Pass rate	The run's pass rate, shown as a bar and a percentage. It appears only for completed runs; other runs show a dash.
Started	The date and time the run started.

Select the arrow (>) at the end of a row to open the run's details. See Run details. Use Rows per page and the page controls at the bottom of the table to move through older runs.

Figure 4. The Recent runs table.

Filter runs

Use the controls above the table to narrow the list:

• Running shows how many runs are in progress right now.
• All targets filters by target type: Agent, Model, Gateway app, or AI workflow.
• All statuses filters by run status.

Figure 5. Filter by target type.

Figure 6. Filter by status.

Run statuses

Status	Meaning
Queued	The run is waiting to start.
Running	The run is in progress.
Completed	The run finished. Its pass rate is shown.
Incomplete	The run ended before all of its attacks finished.
Failed	The run could not finish, for example because the target could not be reached. This is a run error, not a test result.
Cancelled	The run was stopped before it finished.