Deception Scanner

This tool analyzes a single input string of text to detect semantic markers of manipulative trust-building, unverified promises, and forced teaming. It accepts one required parameter text, which is the raw content to analyze and is treated as authoritative input. The worker infers a deception probability score, supporting rationale, and specific flagged phrases or spans, as defined by the system-provided structured output schema. The tool output must strictly validate against that system-provided structured output schema (auto-produced, not supplied here).