Enable Guardian for External Agents
Use a system property to protect your Language Learning Models (LLMs) with Guardian.
Use the sn_aia.external_agent_guardian_check system property to enable Guardian, which monitors requests sent to Language Learning Models (LLMs) and their responses to help protect you, your users, and your data. There are three types of content that are monitored for:
- offensive or harmful content
- prompt injection attempts
- filtered subjects
Ensure that the sn_aia.external_agent_guardian_check system property does not exist in the System Properties [sys_properties] table, or exists and is set to a value of true.
More information
| Attribute | Description |
|---|---|
| Configuration name | sn_aia.external_agent_guardian_check |
| Configuration type | System Properties \(/sys\_properties\_list.do\) |
| Data type | Boolean |
| Recommended value | true |
| Default value | true |
| Fallback value | true |
| Category | Access control |
| Security risk | - Severity score:4.1 - CVSS score: Medium - Security risk details: Prompt injection is a type of attack where bad actors override the normal instructions of an LLM to access restricted information or elicit unexpected behaviors. Prompt injection detection is based on the LLM which has been trained on various types of prompt injection techniques such as role playing, paraphrasing, repetition, instructions to ignore other instructions, persuasion, etc. |
| Functional Impact | Guardian may produce false positives and block legitimate requests. |
| Dependencies and prerequisites | None |
Parent Topic:Access control