feat(bot): answer treatment questions from the clinical-guideline knowledge base

- menu button «Вопрос о лечении» -> kb.question -> rag.answer (KR quote + source)
  or «уточните у врача»; dose questions get a fixed doctor redirect
- free text in idle that looks like a question (no check-in signals, not a drug
  question) -> knowledge base; no answer -> the usual support phrase
- routing order unchanged: consent/safety first, side effects/quit/alcohol -> check-in,
  own drug -> med_qa (instruction only, SPEC §7 p.5)
- question cues, small talk and texts in config/texts/knowledge.yaml; scrub placeholders
  before "что такое X" no longer hide the definition question

Refs #64

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

app/llm/rag/qa.py

@@ -56,6 +56,7 @@ PRESCRIBING = re.compile(
 )
 # "что такое флэшбек" -> the term; its definition "Флэшбек (flashback – англ.) – ..." is preferred.
 DEFINITION_QUESTION = re.compile(r"^\s*что (?:такое|значит|означает|это за)\s+(.+?)[?.!\s]*$", re.IGNORECASE)
+LEADING_PLACEHOLDERS = re.compile(r"^(?:\s*\[[A-Z_]+\][\s,.:;!-]*)+")  # "[NAME], что такое ПТСР?"
 DEFINITION_HEAD = re.compile(r"\s[–—-]\s(?![^(]*\))|\s[–—-]?\s?это\s")  # dash outside "(… – англ.)", or "это"
 DEFINITION_HEAD_CHARS = 60
 DEFINITION_BONUS = 2.0  # multiplier for a sentence that starts with the asked term (picks it within a chunk)
@@ -174,7 +175,7 @@ def _allowed(sentence: str | None, audience: Audience) -> bool:
 
 
 def _defined_term(question: str) -> set[str] | None:
-    m = DEFINITION_QUESTION.match(question)
+    m = DEFINITION_QUESTION.match(LEADING_PLACEHOLDERS.sub("", question))
     return set(tokens(m.group(1))) or None if m else None