Grandmas Wisdom

Grandmas Wisdom is a skill for AI agents that evaluates whether academic citations actually support the claims being made from them. Rather than performing a simple cross-reference check, it analyzes evidential substance, logical consistency, and longitudinal validity of scholarly sources. The core output is a 1–10 Bullshit Meter score with an actionable interpretation, giving an agent a concrete signal about whether a citation is trustworthy. Alongside the score, the skill produces a structured breakdown of evidential support, flags overclaim risks, and delivers explicit Tenable or Not Tenable verdicts for each claim under review. It also performs longitudinal reevaluation as new literature is discovered, meaning a claim that was defensible earlier can be revisited when the evidence base changes. Internally the skill uses a multi-pass process that maps conceptual connections and measures the gap between what a source literally supports and what is being inferred from it. This is distinct from asking a general-purpose model to judge truthfulness, because the skill applies specialized reasoning structures including what the description calls a Spiral Reasoning Tree and Veritas Aegis safeguards. The intended use case is research workflows where an AI agent must produce defensible academic work and cannot afford to propagate hallucinated or miscited sources. There are no environment variables or external tool integrations listed for this skill.