fix: 6 audit issues — missing await, regime range, pct_out scaling, ticker validation, dead code, flag merge

1. app.py: await _update_in_progress (coroutine was silently dropped)
2. models.py + tier1.py: regime_score_adjustment range ±2→±10 (was negligible on 0-100 scale)
3. y_finance.py: pct_out * 100 (was fraction, displayed as percent)
4. app.py: ticker validation accepts dots/hyphens (BRK.B, BF-B)
5. portfolio.py: wire _fetch_peer_basics into theme substitution (was dead code)
6. setup.py: accumulate global_flags across parallel agents (dict.update was dropping them)

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

app.py

@@ -1,6 +1,7 @@
 """FastAPI SSE backend for the structured equity ranking engine."""
 
 import os
+import re
 import time
 import uuid
 import asyncio
@@ -231,7 +232,7 @@ async def _run_analysis_inner(analysis_id: str, ticker: str, trade_date: str):
                     await q.put(evt)
 
             # Mark in-progress agents for upcoming stages
-            _update_in_progress(chunk, emitted_fields, prev_agent_statuses, state, q, start_time)
+            await _update_in_progress(chunk, emitted_fields, prev_agent_statuses, state, q, start_time)
 
     except Exception as e:
         print(f"[ANALYSIS] Stream error: {e}\n{_tb.format_exc()}", flush=True)
@@ -396,7 +397,7 @@ async def _start_cleanup():
 @app.post("/analyze", dependencies=[Depends(verify_api_key)])
 async def start_analysis(req: AnalyzeRequest):
     ticker = req.ticker.upper().strip()
-    if not ticker or len(ticker) > 5 or not ticker.isalpha():
+    if not ticker or not re.match(r'^[A-Z0-9.\-]{1,6}$', ticker):
         raise HTTPException(400, "Invalid ticker")
     trade_date = req.date or str(date.today())
     analysis_id = str(uuid.uuid4())

tradingagents/agents/structured/portfolio.py

@@ -104,14 +104,36 @@ def create_theme_substitution_node(llm):
             industry = card.get("industry", "")
             sector = card.get("sector", "")
 
-        # Fetch peer data for comparison
-        # First, ask LLM to identify theme peers, then we'll fetch their data
+        # Fetch competitor/peer data to ground the LLM's comparison
+        competitors = card.get("competitors") or []
+        peer_data = _fetch_peer_basics(competitors) if competitors else []
+        peer_summary = ""
+        if peer_data:
+            lines = []
+            for p in peer_data:
+                if p.get("error"):
+                    continue
+                rg = p.get("revenue_growth")
+                rg_str = f"{rg*100:.1f}%" if rg else "N/A"
+                pm = p.get("profit_margins")
+                pm_str = f"{pm*100:.1f}%" if pm else "N/A"
+                lines.append(
+                    f"  {p['ticker']}: P/E={p.get('trailing_pe', 'N/A')}, "
+                    f"Fwd P/E={p.get('forward_pe', 'N/A')}, "
+                    f"RevGrowth={rg_str}, "
+                    f"Margins={pm_str}, "
+                    f"52W={p.get('52w_range_pct', 'N/A')}%"
+                )
+            peer_summary = "\n".join(lines)
+
         theme_prompt = f"""You are a Theme Substitution Analyst. Your job: determine if {ticker} is the BEST
 expression of its investment theme, or if better alternatives exist.
 
 CANDIDATE STOCK:
 {summary}
 
+{f'PEER FUNDAMENTALS (live data):{chr(10)}{peer_summary}' if peer_summary else 'No live peer data available — use your knowledge of these companies.'}
+
 INSTRUCTIONS — do this in order:
 
 1. IDENTIFY THE THEME: What macro/sector theme does {ticker} express?
@@ -120,9 +142,10 @@ INSTRUCTIONS — do this in order:
    Name it clearly in theme_name.
 
 2. LIST THEME PEERS: Name 3-6 other publicly traded stocks that express the SAME theme.
-   These should be the strongest competitors for capital allocation in this theme.
-   For each peer, estimate a master_score_estimate (0-10) based on your knowledge of
-   their fundamentals, momentum, and positioning vs {ticker}.
+   Use the peer data above if available. These should be the strongest competitors
+   for capital allocation in this theme.
+   For each peer, score master_score_estimate (0-10) based on fundamentals, momentum,
+   and positioning vs {ticker}.
 
 3. RANK WITHIN THEME: Rank all stocks (including {ticker}) by investment quality.
    The stock with the best combination of: business quality, valuation, momentum,

tradingagents/agents/structured/tier1.py

@@ -188,13 +188,13 @@ INSTRUCTIONS:
 2. Classify liquidity_regime: "expansion" / "contraction" / "neutral".
    - expansion: falling yields, dovish Fed, credit flowing, dollar weakening.
    - contraction: rising yields, hawkish Fed, tight credit, dollar strengthening.
-3. Set regime_score_adjustment (-2 to +2):
-   - +2 = strong macro tailwind for this specific stock/sector.
-   - +1 = mild tailwind.
+3. Set regime_score_adjustment (-10 to +10):
+   - +5 to +10 = strong macro tailwind for this specific stock/sector.
+   - +1 to +4 = mild tailwind.
    -  0 = neutral.
-   - -1 = mild headwind.
-   - -2 = severe macro headwind (risk-off + contraction + hostile sector).
-   This adjustment directly modifies the master score for ALL stocks.
+   - -1 to -4 = mild headwind.
+   - -5 to -10 = severe macro headwind (risk-off + contraction + hostile sector).
+   This adjustment directly modifies the 0-100 master score for ALL stocks.
 4. Score macro_alignment_0_to_10: how well macro supports {ticker} specifically.
 5. Also provide score_0_to_10 (overall macro health).
 6. Set regime_label: descriptive label (e.g., "Late Cycle Risk-Off").

tradingagents/dataflows/y_finance.py

@@ -613,7 +613,7 @@ def get_institutional_flow(ticker):
                     {
                         "holder": str(r.get("Holder", "")),
                         "shares": int(r["Shares"]) if r.get("Shares") else None,
-                        "pct_out": float(r["% Out"]) if r.get("% Out") else None,
+                        "pct_out": round(float(r["% Out"]) * 100, 2) if r.get("% Out") else None,
                         "value": float(r["Value"]) if r.get("Value") else None,
                     }
                     for r in top

tradingagents/graph/setup.py

@@ -191,11 +191,16 @@ def _create_parallel_node(agent_fns: List[tuple], label: str):
         results = await asyncio.gather(*tasks, return_exceptions=True)
 
         merged: Dict[str, Any] = {}
+        all_flags: list = []
         for (name, _), result in zip(agent_fns, results):
             if isinstance(result, Exception):
                 logger.error("[%s] %s failed: %s", label, name, result)
                 continue
+            flags = result.pop("global_flags", [])
+            all_flags.extend(flags)
             merged.update(result)
+        if all_flags:
+            merged["global_flags"] = all_flags
 
         logger.info("[%s] completed in %.1fs", label, time.time() - t0)
         return merged

tradingagents/models.py

@@ -121,9 +121,9 @@ class MacroRegimeOutput(AgentBaseOutput):
     risk_appetite: str = "neutral"  # risk-on / risk-off / transitional
     liquidity_regime: str = "neutral"  # expansion / contraction / neutral
     regime_score_adjustment: float = Field(
-        default=0.0, ge=-2, le=2,
-        description="Adjustment applied to all downstream scores. "
-                    "+2 = strong macro tailwind, -2 = severe macro headwind.",
+        default=0.0, ge=-10, le=10,
+        description="Adjustment applied to the 0-100 master score. "
+                    "+10 = strong macro tailwind, -10 = severe macro headwind.",
     )