probability-confidence-calibration

What problem does it solve?

This Skill helps you correct overconfidence and underconfidence by aligning your stated certainty with the actual strength and completeness of the evidence behind a claim.

Core Features & Use Cases

• Confidence calibration workflow: Turns a claim with a stated percentage into an evidence-audited, frequency-tested calibrated estimate.
• Evidence strength auditing: Separates support into observation, inference, anecdote, and assumption to expose weak foundations.
• Counter-evidence and failure-mode weighting: Surfaces what would be expected if the claim were true, identifies missing checks, and reweights confidence around the most likely way the claim fails.
• Use case: When you’re asked “How sure are you?” after making a judgment (e.g., research interpretation, risk assessment, or decision-making), this skill produces a more defensible calibrated confidence level.

Quick Start

Calibrate your confidence for the claim “X is true” with an original confidence of 80% by listing supporting evidence, counter-evidence, the most likely failure mode, and then returning a calibrated confidence percentage.