Struct PromptGuard 

pub struct PromptGuard { /* private fields */ }

The prompt injection detection engine.

Maintains a configurable set of detection rules and a severity threshold for blocking. Inputs that match patterns at or above the threshold are blocked; those with lower-severity matches produce warnings.

Implementations

impl PromptGuard

pub fn new() -> Self

Create a new guard with built-in detection patterns and a default block threshold of High.

pub fn with_config(config: PromptGuardConfig) -> Self

Create a guard with custom configuration.

pub fn set_block_threshold(&mut self, threshold: InjectionSeverity)

Set the minimum severity level that triggers blocking.

pub fn config(&self) -> &PromptGuardConfig

Get the current configuration.

pub fn add_pattern( &mut self, name: &str, pattern: &str, severity: InjectionSeverity, description: &str, )

Add a custom detection pattern.

pub fn scan_input(&self, text: &str) -> Vec<InjectionAlert>

Scan input text and return all injection alerts (pattern matching only).

pub fn scan(&self, input: &str) -> PromptGuardResult

Full scan with scoring, structural analysis, and threat assessment.

pub fn is_safe(&self, input: &str) -> bool

Quick check: returns true if the input is considered safe.

pub fn sanitize(&self, input: &str) -> String

Sanitize input by stripping detected injection patterns.

pub fn scan_and_decide(&self, text: &str) -> ScanDecision

Scan input text and return a decision: Allow, Warn, or Block. This is the legacy API; prefer scan() for richer results.

Trait Implementations

impl Clone for PromptGuard

fn clone(&self) -> PromptGuard

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for PromptGuard

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Default for PromptGuard

fn default() -> Self

Returns the “default value” for a type.