@penguinharness/continual-learning
v0.2.11
Published
Continual learning: a stop hook that, when the task that just ended ran more than 30 completed turns, hands a condensed excerpt of that task to a background subagent, which folds the durable findings into the agent's own skills.
