fix: correct context pruning order and vector alignment

- Prune context BEFORE processing (LSTM forget gate pattern)
  so LLM only sees relevant context, instead of pruning after the fact
- Fix ensureTrained() to recompute all event vectors after retraining
  vectorizer, fixing feature-space mismatch between stored vectors and
  query vector that made relevance scoring effectively random
- Add read lock to knowledge BuildTree() (data race fix)
- Log writeIndex() errors instead of discarding them
- Fix TOCTOU race in document ContextToDoc() dedup
- Fix healthcheck timing (measure elapsed before cleanup)
- Refactor waiter CLI into separate files (state, conn, config, editor,
  history, builtin) for maintainability
- Add CLI plugin API key authentication

internal/agent/core/context.go

@@ -223,6 +223,10 @@ func (c *RelevanceContext) ensureTrained() {
 			texts[i] = evt.Input + " " + evt.Response
 		}
 		c.veczer.Train(texts)
+		// 重算所有事件向量，与新的向量化器特征空间对齐
+		for _, evt := range c.events {
+			evt.Vector = c.veczer.Vectorize(evt.Input + " " + evt.Response)
+		}
 		c.trained = true
 	}
 }