Conversational social engineering detection using a Finite State Machine.

No LLMs. No embeddings. No semantic search. Pure algorithmic detection.

How it works

xtra models conversations as state transitions:

NEUTRAL → RAPPORT_BUILD → TRUST_ESTABLISHED → EXTRACTION_ATTEMPT → FLAGGED

Attacks are detected by their path through state space, not their content.

Signals tracked

• Flattery density and momentum across turns
• Asymmetry spike (give/ask ratio collapse)
• Escalation velocity
• Reciprocity pressure
• Decoy turn detection
• Scope mismatch

Usage

from xtra import Xtra, TurnSignals

result = Xtra().analyze(turns)
if result.is_attack:
    print(result.flag_reason, result.confidence)

Part of Tangible Research

Making AI Tangible — github.com/TangibleResearch