evaluate-shortform

A pattern log is a canonical, falsifiable 4-line block that records evaluation results for downstream tools to diff and reason about. It enables cross-cycle learning by providing a standardized format for longitudinal analysis of short-form posts.

Run the evaluation with the loop slug, post URL, and brief path to generate a cycle artifact. This process runs after publishing or during an eval cycle, using the loop workspace as context to write results for traceability.

Yes, the evaluation runs within an existing loop and uses the loop workspace as context. It writes results to the loop's evals workspace, providing traceability and enabling longitudinal analysis of short-form posts across cycles.

Short-form evaluation runs during an eval cycle within an existing loop, using the loop workspace as context. It scores the published post against the brief and platform catalog, then logs results to the loop's evals workspace for longitudinal analysis.

A pattern log is a canonical 4-line block that records evaluation results for downstream tools to diff and reason about. It is needed to enable cross-cycle learning and longitudinal analysis by providing a falsifiable, standardized entry for each short-form post.