Galileo Experimental Prover MCP Connector for Claude

Galileo did not accept 2,000 years of Aristotelian physics — he tested it. You must: (1) QUESTION AUTHORITY — challenge the claim at its source. Evaluate evidence, not prestige. "The documentation says" is a starting hypothesis, not proof. "I tested the documented behavior and confirmed/denied it" is evidence, (2) DESIGN an EXPERIMENT — create a specific, repeatable test with hypothesis, method, measurement instrument, and repeatability protocol. Galileo built his own telescope and water clock because existing instruments were inadequate, (3) CONTROL VARIABLES — change ONLY ONE variable, hold everything else constant. If you change two things and the result improves, you do not know which caused it, (4) PREDICT OUTCOME before testing — state what you expect, what confirms, what denies, and what single result would FALSIFY the hypothesis entirely. Without a prediction, any result confirms bias, (5) REVISE BELIEFS when evidence contradicts them — update your position, do not dismiss the data. The Aristotelians refused to look through the telescope. Do not be them. If rejected, your reasoning is faith-based, not evidence-based.Send ALL parameters in ONE call — the analyses above plus YOUR OWN verdict: every PIVOT boolean, verdict and clarification. The engine cross-checks the pivots against the verdict and rejects incomplete or self-contradictory calls. Structured reflection tool for Galileo-level experimental reasoning — forces questioning authority with evidence (not prestige), designing controlled experiments with measurable predictions, isolating single variables, stating predictions BEFORE testing, and revising beliefs when evidence contradicts them. Catches Authority Deferred (accepting a claim because of WHO said it instead of WHAT evidence supports it — "the vendor whitepaper says 10x performance" is deference. "I benchmarked the vendor's tool against our current solution on our production dataset and measured 2.3x improvement, not 10x" is evidence. Galileo did not accept Aristotle's 2,000-year-old claim that heavy objects fall faster — he dropped two masses from Pisa and measured), Experiment Absent (proposing solutions without testing them — "it stands to reason" is not evidence. A specific hypothesis, a defined method, a measurement instrument with known precision, and a repeatable protocol IS an experiment. Galileo built his own 20x telescope and water clock to 1/100th second precision because existing instruments were not accurate enough), Variables Uncontrolled (changing multiple things at once and attributing results to one — "we deployed the new code AND upgraded the database AND added caching, and performance improved." Which change caused the improvement? Galileo's inclined plane: changed ONLY the angle, same ball, same surface, same timing), Prediction Missing ("let us see what happens" instead of stating the expected outcome BEFORE running the test — without a prediction, any result confirms your hypothesis. Galileo predicted equal fall rates BEFORE dropping the masses), and Dogma Persistent (dismissing contradicting evidence as "anomalous" instead of updating beliefs — the Aristotelians REFUSED to look through Galileo's telescope because the evidence would challenge their worldview. Do not refuse to look through the telescope). Call once per claim validation, A/B test, or experimental design