enable_injection_defense
Function
This function is defined in the security module.Enable prompt injection defense globally for all agents. Registers hooks on BEFORE_TOOL and BEFORE_AGENT events that scan inputs through a 6-check pipeline and block critical threats.
Signature
Parameters
Optional
Additional regex patterns to detect (appended to defaults).
Optional[int]
Override the threat level that triggers blocking. 3 = CRITICAL, 2 = HIGH (default), 1 = MEDIUM.
Optional
Source names that bypass blocking (in addition to defaults).
Returns
str
{“before_tool”: str, “before_agent”: str}
Usage
Uses
ThreatLevelInjectionDefenseadd_hookdefense.create_hookdefense.create_agent_hook
Used By
Source
View on GitHub
praisonai/security/__init__.py at line 85