Confidence Scorer Skill

What problem does it solve?

This Skill resolves uncertainty in detected technology stacks by converting corroboration signals into consistent High, Medium, or Low confidence labels.

Core Features & Use Cases

It scores each correlated technology using weighted evidence signals such as source diversity, signal strength, cross-source agreement, evidence type preference (technical vs job postings), and conflict penalties. It also produces actionable outputs including a per-technology confidence result with detailed reasoning, an aggregate confidence summary across a report, and a list of confidence gaps that indicate where additional validation signals are needed. Typical use cases include technology stack identification from multi-source analysis where findings must be prioritized for follow-up verification.

Quick Start

Use the confidence_scorer operation to score each correlated technology from the correlated_technologies array and produce confidence-labeled results with explanations.