Update src/api/agent.py

src/api/agent.py

@@ -1575,6 +1575,7 @@ def _light_strip_retrieval_boilerplate(text: str) -> str:
     cleaned = re.sub(r"^\s*please\s+", "", cleaned, flags=re.IGNORECASE)
     cleaned = re.sub(r"\s+(please|thanks|thank you|asap)$", "", cleaned, flags=re.IGNORECASE)
     # Strip search intent verbs — these are action signals, not topic content
+    # Pass 1: strip leading intent verb phrases (including compound forms)
     cleaned = re.sub(
         r"^\s*(find|search for|search|look for|get me|show me|give me|fetch|retrieve|"
         r"research on|articles on|papers on|literature on|studies on|"
@@ -1583,6 +1584,18 @@ def _light_strip_retrieval_boilerplate(text: str) -> str:
         r"i want articles on|i need papers on|i want papers on)\s+",
         "", cleaned, flags=re.IGNORECASE
     )
+    # Pass 2: strip residual noun boilerplate left after pass 1
+    # e.g. "find research on 33" → pass1 → "research on 33" → pass2 → "33"
+    cleaned = re.sub(
+        r"^\s*(research|articles?|papers?|books?|literature|studies|study|"
+        r"journals?|publications?|resources?)\s+(on|about|for|regarding|into)\s+",
+        "", cleaned, flags=re.IGNORECASE
+    )
+    # Pass 3: strip dangling prepositions after intent removal
+    cleaned = re.sub(
+        r"^\s*(on|about|for|regarding|concerning|into|around|of|in|the)\s+",
+        "", cleaned, flags=re.IGNORECASE
+    )
     return re.sub(r"\s+", " ", cleaned).strip()
 
 
@@ -1751,16 +1764,18 @@ async def _build_search_plan(question: str, model: str) -> dict:
     light_query = _light_strip_retrieval_boilerplate(raw_query) or raw_query
 
     # ── Ambiguity gate ────────────────────────────────────────────────────────
-    # After stripping intent verbs, check if what remains is meaningful enough
-    # to build a boolean query. Numbers, single characters, and very short tokens
-    # cannot produce a useful search and should prompt clarification.
+    # After stripping intent verbs and prepositions, check if what remains
+    # is meaningful enough to build a boolean query.
     _stripped = light_query.strip()
     _is_pure_number  = bool(re.fullmatch(r"[\d\s\.\,\-\/]+", _stripped))
     _is_too_short    = len(_stripped) < 2
     _is_single_char  = bool(re.fullmatch(r"[a-zA-Z0-9]", _stripped))
+    # Check if remaining words are all numbers/single chars — no real content words
+    _content_words   = [w for w in _stripped.split() if len(w) > 1 and not re.fullmatch(r"[\d\.\,\-]+", w)]
+    _no_content      = len(_content_words) == 0 and len(_stripped) > 0
 
-    if _is_pure_number or _is_too_short or _is_single_char:
-        _display = _escaped = _stripped or raw_query
+    if _is_pure_number or _is_too_short or _is_single_char or _no_content:
+        _display = _stripped or raw_query
         return {
             "corrected":            raw_query,
             "natural":              light_query,