Learning Signals from Trajectories

The starting point of evolution is not generalization, but evaluation. A three-level check: outcome checker (tests, DB state – “is the task actually done?”), process checker (rules, permissions – “done in an allowed way?”), quality checker (LLM rubric – “done well?”). The lower the level, the more reliance on code; the language model is only for the hard-to-formalize. Correct outcome ≠ correct process: deleting failing tests also gives green tests. The evaluation result is a structured diagnosis, not a scalar. The checker itself needs calibration.

Related: [Continuous Agent Evolution], [Four Principles of Rubric], [Four Carriers of Update]