Viewing and Enabling Policies
UnifAI provides a centralized interface that allows you to view, enable, and upload policies. To view UnifAI's out-of-the-box policies, log in to the UnifAI dashboard. You can also browse Commonly Searched Policies. For a complete list of policy related prompts you can use, see What Can I Ask UnifAI?
In UnifAI, you can accomplish the following tasks related to your policies:
View all policies
Enter View all policies prompt to see all policies.
The page displays the total number of policies, including UnifAI’s out of the box policies and your custom policies. These policies are grouped by category and asset type for easier analysis.
On the View All Policies page:
Lineaje Policies: Shows the total number of Lineaje policies and their status.
Custom Policies: Shows the custom policies you imported and their status.
Policies by Category: Shows enabled vs. disabled policies across the categories of AI Threats and Exploits, Data Security and Privacy, Identity and Access Control, and Vulnerability, and Compliance.
Assets by Type: Shows enabled vs. disabled policies across the asset types.
In the All Policies section, expand each category for more details:
Lineaje Policies:
AI Asset Type: The AI asset the policy applies to, such as an AI agent, model, skills, or MCP server.
Compliance Standards: Indicates the regulatory or industry standard the policy maps to, like EU AI Act or OWASP‑LLM.
Policy: Specifies the required security control for the AI asset.
Policy Status: Shows whether the policy is enabled or disabled.
Custom Policies: Shows your policies that you have imported into UnifAI.
Enable or disable policies
In the Policy Status column, use the toggle to enable or disable policies for your assets.
When the toggle is set to
Enabled, all associated controls are actively evaluated and enforced.When the toggle is set to
Disabled, policy evaluation is halted and no controls are applied to the assets.
Upload custom policies
Custom policies allow you to extend UnifAI's out-of-the-box controls with your organization-specific requirements.
1. Enter upload policies or create custom policies prompt to upload your custom policies.
You can upload your custom policies in PDF format with a maximum size of 10MB.
After the policies are extracted, you can review and edit them prior to submission.
After submission, Lineaje team reviews the policies and notifies you when implementation is complete.
Your custom policies appear in the All Policies section. Enter prompt
See all policiesto see your custom policies.
Commonly Searched Policies
Explore the high‑impact policies that organizations most often rely on to enforce AI security standards and maintain compliance:
Allow/Block
Do not use LLMs from the organization's disallowed list.
Only use those LLMs that are in the organization’s approved list.
Prompt Injection
Do not allow malicious content via prompts included in uploaded files.
Inter agent communication must be authenticated.
OWASP Top 10 and EU AI Act
Do not expose PII on the UI.
No file should contain any PII.
OWASP Top 10 and NIST SSDF
Do not allow dependencies with critical or high severity vulnerabilities.
Do not allow critical or high vulnerabilities in the code.
Skills
Do not allow use of malicious skills
Do not allow use of suspicious skills
For a complete list of UnifAI policies, log in to the UnifAI dashboard.
Last updated