ReadonlycategoryWhich trap category this detector addresses
ReadonlyidUnique identifier for this detector
ReadonlynameHuman-readable name
Protected ReadonlytrapReturn sanitized content with threats neutralized
Scan content and return any threats found
Optionaloptions: DetectorOptions
Detects sub-agent spawning traps.
Reference: Triedman et al. (2025) — adversarial content hijacks multi-agent control flow with 58-90% success rate.