review-hypotheses

Track prediction accuracy by resolving active hypotheses, learning from outcomes, and calculating accuracy metrics. The skill records results and updates internal models to improve future predictive capabilities.

Hypothesis testing in continuous learning detects outcomes of active predictions, extracts patterns from results, and calibrates confidence. This systematic resolution updates the agent's memory tree to enhance ongoing accuracy.

Resolve active hypotheses automatically using script-driven resolution to detect outcomes and record results. This deterministic execution updates the memory tree and processes resolved records for reflective learning.

Generate accuracy reports by processing resolved hypotheses to calculate detailed statistics across various dimensions. The skill tracks violations and calibrates confidence to provide comprehensive performance reporting.

Structured pipeline records are required to perform outcome analysis, as the skill operates on structured hypothesis records within a defined pipeline to verify predictions and extract learning patterns.

Inaccurate predictions occur when outcomes are not resolved or learned from. The skill processes resolved hypotheses to extract patterns, track violations, and calibrate confidence, directly improving future prediction accuracy.