world-model-principles-scorecard

A world model health scorecard audits whether a knowledge system compounds value or rots. You need one when monitoring long-term reliability in agent memory, knowledge bases, or decision pipelines to prevent stale or contradictory knowledge.

You audit knowledge rot by evaluating the system against five principles: signal fidelity, earned structure, outcome encoding, organizational resistance, and accumulated reality. This generates a 25-point scorecard identifying gaps and improvement priorities.

Yes, the scorecard evaluates decision pipelines and knowledge bases by scoring five core principles to generate actionable improvement priorities and trajectory insights, monitoring long-term reliability and alignment.

The 25-point scorecard output includes per-principle evidence, gaps, risk levels, and concrete improvement steps. It provides trajectory insights to determine whether your world model is improving in fidelity or drifting toward stale knowledge.

You assess signal fidelity by scoring the system's world model against established principles as part of the 25-point scorecard. This evaluates whether accumulated reality and outcome encoding maintain high data quality over time.

You should avoid using this scorecard for systems lacking persistent memory or accumulated knowledge structures. It specifically audits long-term knowledge reliability and organizational resistance, making it less suited for stateless or ephemeral data processing pipelines.