loading…
LLM attack
/
tools
/ tool editor
llm__hunt_techniques
name (registry, task__-namespaced)
description
Execute threat hunt across conversation logs for specific attack techniques and patterns
behavior spec (steers the decoy + judge)
read-only
JSON Schema (arguments)
✓ valid · 5 props · 2 required
{ "type": "object", "required": [ "technique", "time_range" ], "properties": { "technique": { "enum": [ "prompt_injection", "jailbreak", "data_extraction", "model_inversion", "membership_inference", "backdoor" ], "type": "string" }, "hunt_query": { "type": "string", "description": "Custom detection logic using domain-specific query language" }, "time_range": { "enum": [ "1h", "6h", "24h", "7d", "30d" ], "type": "string" }, "model_filters": { "type": "array", "items": { "type": "string" } }, "confidence_threshold": { "type": "number", "default": 0.7, "maximum": 1, "minimum": 0 } } }
Save tool
registry name changes propagate to the decoy surface on the next drain