Enforcement Rules (Early Availability)
Enforcement rules define which AI activity AI Exposure blocks. When Anthropic sends a Claude prompt to the Tenable endpoint, Tenable One AI Exposure checks it against your enabled rules and returns an allow or deny verdict. A denied prompt reaches the Claude platform, but never reaches the model and the blocked prompt appears in Tenable One AI Exposure as an activity detection. Create rules to match the activity types, severities, and content types you want to stop.
For information about configuring inference hooks in Claude, see Policy Enforcement for Claude (Early Availability)
Enforcement Rules Column Data
The following columns appear in the table on the Enforcement Rules page.
| Filter | Description |
|---|---|
| Name | The name of the enforcement rule. |
| Description | A description of the enforcement rule. |
| Activity Rules | The list of activity rules, either all activity rules or specific activity rules. |
| Severity | The severity level of the policy, which determines the severity level of any findings the policy triggers. The default is Critical and High. |
| Content Types | The list of content types for the selected activity rules. |
| Status | The status of the enforcement rule (enabled or disabled). Only enabled rules block prompts. Disables rule do not trigger enforcement actions. |
| Modified By | The user who modified the enforcement rule. |
| Modified Time | The date and time when someone last modified the enforcement rule. |
Create an Enforcement Rule
An enforcement rule determines which AI activityTenable One AI Exposure blocks.
To create an enforcement rule:
-
Go to Policies > Enforcement Rules .
The Create Enforcement Rule page appears.
-
Click Add Enforcement Rule.
The Create Enforcement Rule page opens.
On the right, AI Activity Detections lists existing activity detections that match the activity rules, severities, and content types you select.
-
In the Name box, type a name for the rule.
-
(Optional) In the Description box, type a description for the rule.
-
In the Activity rules section, select one of the following options:
-
All activity rules
-
Specific activity rules: Click + Add and select at least one activity rule.
-
-
In the Severity section, select one of the following:
-
All severities
-
Specific severities: Select one or more severity levels from the list. Default is Critical and High.
-
-
In the Content types section, select one of the following options:
-
All content types
-
Specific content types: Click +Add and select at least one content type.
-
-
In the Block message box,type a message to display whenTenable One AI Exposure blocks a prompt. This box supports up to 500 characters. If you leave this box empty, the system displays the following default message::
Content blocked by Tenable One AI Exposure (content type) -
Click Create.
Tenable One AI Exposure saves the enforcement rule.
Enable and Disable Enforcement Rules
No enforcement rules exist by default. Until you create and enable a rule, Tenable One AI Exposure does not block any prompts. Disabling or deleting a rule stops it from blocking new prompts, though existing detections remain.
To enable or disable an enforcement rule:
- In the AI Exposure Console, click Policies > Enforcement Rules. to view all rules.
The table displays the following columns: Name, Description, Activity Rules, Severity, Content Types, and Status (Enabled or Disabled).
-
(Optional) Use the filters to narrow the list.
-
Locate the enforcement rule you want to update.
- Click the menu
button and click Enable or Disable. - Click Save.
Manage Existing Enforcement Rules
You can edit, duplicate, or delete an existing enforcement rule.
To manage an existing rule:
- In the AI Exposure Console, click Policies > Enforcement Rules.
The table displays the following columns: Name, Description, Activity Rules, Severity, Content Types, and Status (Enabled or Disabled).
-
(Optional) Use the filters to narrow the list.
- Locate the enforcement rule you want to manage.
- Click the menu
button and select one of the following options:- Edit
- Delete
- Duplicate
- Click Save.
Review Blocked Activity
Before you Begin
-
To view the AI Enforcement message, you must have the AI Conversation Viewer role. For more information, see AI Conversations.
To review blocked activity:
-
In AI Exposure, go to Risks > Activity Detections.
Tenable One AI Exposure lists all detections. The Action column displays Blocked or Detected Only. Blocked detections include a block icon. -
Open a blocked detection to view the following:
-
Enforcement Rule: The rule that blocked the prompt, including a link to the rule.
-
Evidence: The detection evidence, followed by a View Message link. Clicking this link opens the AI Enforcement Message, which displays the blocked prompt with the identity and model name.
Note: In shadow mode, Claude does not block any prompts. Use activity detections in shadow mode to verify your rules before switching to block mode.
-