glpi-neuroforge-mega/RELEASE-NOTES-v1.4.1.md

GLPI NeuroForge Mega v1.4.1

Fix: Goal learning under Controlled Learning

• Adds NEUROFORGE_GOAL_LEARNING_ENABLED as an explicit, independent gate.
• Keeps the safe default false.
• Allows goal-cycle learning while NEUROFORGE_CONTROLLED_LEARNING=true remains enabled.
• Does not re-enable raw chat input learning, assistant-response learning, or imports.
• Wires the flag through Docker Compose and exposes it read-only in Control Center config.