What problem does it solve?
In an era of information overload, misleading claims, selective reporting, and unstated cognitive biases make it difficult to distinguish verified facts from framed narratives or unsubstantiated assertions. This Skill eliminates the guesswork of evaluating content credibility by providing a structured, evidence-based deep fact-checking workflow that uncovers hidden manipulation and quantifies uncertainty.
Core Features & Use Cases
- Three-Layer Fact Model: Separates empirical facts (L1), interpretive frames (L2), and value judgments (L3) to apply the right verification strategy for each claim type.
- Manipulation Detection: Identifies 7 common information manipulation tactics including cherry-picking, temporal manipulation, false equivalence, and authority misattribution.
- Competitive Hypothesis Testing: Generates and tests multiple competing explanations for claims, rather than only seeking confirming evidence, to avoid confirmation bias.
- Bayesian Credibility Scoring: Assigns probabilistic confidence ratings to claims based on source quality and evidence strength, with clear labels for epistemic uncertainty.
- Author Bias Analysis: Detects selective reporting patterns, hidden agendas, and language cues that reveal author predispositions.
Use cases include verifying high-impact public claims, evaluating controversial AI industry predictions, assessing media content for accuracy, and supporting journalistic or research due diligence.
Quick Start
Use the deep-fact-check skill to verify the credibility of the article at https://example.com/ai-industry-prediction and generate a full structured fact-check report with bias analysis and confidence scores.
