---

name: sr3_recovery_evaluation description: Evaluate SR3-style proxy runs with consistency metrics, source-boundary records, and mechanism checks. This reusable skill supports bounded SR3 recovery experiments.

sr3_recovery_evaluation

Use this skill for SR3 proxy recovery when implementing recovery_evaluation. Inputs are tiny numeric arrays or metadata; outputs are deterministic logs or metrics. Workflow follows SR3 paper equations for degradation, noising, denoising loss, reverse refinement, or evaluation as appropriate. Validate with validate_skill_tree.py --run-tests. Limitations: not a full SR3 U-Net implementation and not a substitute for FFHQ/ImageNet-scale training.