Guardrails
Policies
A set of detectors defined and parametrized for specific use cases.
Detectors
Models specialized in detecting specific types of content.
ModernGuard
modern-guard
Detects prompt injection vulnerabilities.
Detector
Security
Prompt Injection
Suspicious Activity Detector
suspicious-session
Identifies unusual patterns indicative of malicious activity based on session data.
Detector
Security
Anomaly Detection
Custom Regular Expression Detector
regex
Detects patterns matching specified regular expressions.
Detector
Pattern Matching
Regex
String Matching Detector
string-match
Identifies occurrences of specific strings within the input.
Detector
String Matching
Keyword Detection
PromptAttack Policy
prompt-attack
Combines prompt injection protection with suspicious activity monitoring.
Policy
PromptAttack
Security