causal_analysis

Yes, you can assess randomized controlled trial results by processing the experimental data or narrative text through this skill. It classifies the evidence type as RCT within a strict hierarchy, evaluates methodological rigor, and outputs a structured JSON object separating robust causal conclusions from mere associations.

To evaluate causal claims from literature reviews, input the narrative text to extract and classify cause-and-effect relationships. The skill applies a strict taxonomy of evidence types including mechanistic and expert opinion, producing a structured JSON report that explicitly separates robust causal conclusions from associations.

Yes, you can use causal inference frameworks to assess the strength of evidence in text by applying counterfactual reasoning to the claims. The skill outputs a structured JSON object detailing claims, evidence type, confounders noted, and an explicit strength assessment based on the evidence hierarchy.

The limitation of causal analysis for mechanistic and expert opinion evidence is that these sit lower in the strict evidence hierarchy than RCTs or observational studies. The skill explicitly classifies these evidence types, evaluates their methodological rigor, and separates their causal conclusions from robust findings in the JSON output.