Struct RAGInjectionGuard 

pub struct RAGInjectionGuard { /* private fields */ }

Guard for detecting hidden instructions in RAG-retrieved documents.

Scans document content for patterns that indicate indirect prompt injection attacks embedded in retrieved context. Pure regex/pattern-based detection.

Requires Professional license.

Implementations

impl RAGInjectionGuard

pub fn new(name: impl Into<String>) -> Result<Self, LicenseError>

Create a new RAG injection guard with default configuration.

License Requirement

RAGInjectionGuard requires a Professional or Enterprise license. Returns an error if the license requirement is not met.

pub fn new_with_license( name: impl Into<String>, license: &LicenseInfo, ) -> Result<Self, LicenseError>

Create a new RAG injection guard with a pre-validated license.

Avoids re-checking the global validator when the caller already holds validated LicenseInfo.

pub fn with_config( name: impl Into<String>, config: RAGInjectionConfig, ) -> Result<Self, LicenseError>

Create with custom configuration and license validation.

Returns an error if the license tier is below Professional.

pub fn with_config_with_license( name: impl Into<String>, config: RAGInjectionConfig, license: &LicenseInfo, ) -> Result<Self, LicenseError>

Create with custom configuration and a pre-validated license.

pub fn with_action(self, action: GuardAction) -> Self

Set the action taken when an injection is detected.

pub fn with_scan_decoded_base64(self, enabled: bool) -> Self

Set whether to scan decoded base64 content.

Trait Implementations

impl Guard for RAGInjectionGuard

fn name(&self) -> &str

Get the guard name

fn check(&self, content: &str) -> GuardCheckResult

Check content against this guard

fn action(&self) -> GuardAction

Get the action this guard takes on failure

fn severity_threshold(&self) -> Severity

Get the minimum severity that triggers this guard