[Pentexa]

…osis

• selection.py: LearningProgressSelector (pooled family signatures, |Δ success| learning progress, per-atom-class weakness targeting, novelty recovery; deterministic, checkpointable)
• diversity.py: Jaccard/entropy/coverage/novelty metrics
• retention.py: frozen-substrate repeated-probe protocol (per-config)
• diagnosis.py: root-causes the v0.1 null result (eval artifact: one-config-per-seed draws; plus trained-state memory saturation)
• curriculum_experiment.py: matched-budget 5-condition comparison (uniform/ascending/calibrated/proc_lp/teacher_lp), per-config eval grids, paired bootstrap CIs, zero API calls, state persistence
• CLI: evren-curriculum, evren-diagnose (warn on hash randomization)
• configs/academy_v01.toml: [evren.curriculum] section
• tests: +5 suites (selection/diversity/retention/diagnosis/experiment)
• docs/EVREN_CURRICULUM_V02.md: protocol + measured results
• pinned run (PYTHONHASHSEED=0): LP and teacher-source effects exactly +0.0000 on classic transfer; ascending only non-uniform (+0.042); retention flat. Honest negative result: student memory saturation is the binding constraint, not curriculum source.

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has insufficient remaining quota. Try requesting a review at a lower effort level.