From bcfaeb7e3930075a4920edfab259c98e344c19f9 Mon Sep 17 00:00:00 2001
From: Jordan Diaz <panksft@hotmail.com>
Date: Fri, 3 Apr 2026 22:56:23 +0000
Subject: [PATCH] =?UTF-8?q?Conversaci=C3=B3n=20continua:=20historial=20com?=
 =?UTF-8?q?o=20mensajes=20user/assistant=20reales?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

El agenticSystem ahora es conversacional — recuerda lo dicho en
mensajes anteriores de la misma sesión.

- engine.py: direct_response guarda en task_history con formato
  "User: X → Agent: Y"
- context/engine.py: _build_messages() reconstruye el task_history
  como pares user/assistant reales en el array de messages, antes
  del mensaje actual. El modelo ve una conversación completa.
- base.py: planner/reviewer no emiten AGENT_DELTA al frontend
  (su output es interno, no para el usuario)

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
---
 src/context/engine.py      | 24 ++++++++++++++++++++++--
 src/orchestrator/engine.py | 17 +++++++++++++++++
 2 files changed, 39 insertions(+), 2 deletions(-)

diff --git a/src/context/engine.py b/src/context/engine.py
index 435d6d1..38d5178 100644
--- a/src/context/engine.py
+++ b/src/context/engine.py
@@ -541,9 +541,29 @@ class ContextEngine:
         else:
             user_content = "Awaiting task assignment."
 
-        messages: list[dict[str, Any]] = [{"role": "user", "content": user_content}]
+        messages: list[dict[str, Any]] = []
 
-        # Append real conversation (assistant messages + tool results)
+        # Include previous task exchanges as conversation history
+        # (so the model remembers what was said in earlier turns)
+        for entry in session.task_history[-10:]:
+            summary = entry.get("summary", "")
+            objective = entry.get("objective", "")
+            if summary.startswith("User: "):
+                # Direct response format: "User: X → Agent: Y"
+                parts = summary.split(" → Agent: ", 1)
+                user_msg = objective or parts[0].replace("User: ", "", 1)
+                agent_msg = parts[1] if len(parts) > 1 else summary
+                messages.append({"role": "user", "content": user_msg})
+                messages.append({"role": "assistant", "content": agent_msg})
+            elif objective:
+                # Task with tools — include as compact exchange
+                messages.append({"role": "user", "content": objective})
+                messages.append({"role": "assistant", "content": summary[:500] if summary else "Tarea completada."})
+
+        # Current user message
+        messages.append({"role": "user", "content": user_content})
+
+        # Append real conversation (assistant messages + tool results from current step)
         if conversation:
             messages.extend(conversation)
 
diff --git a/src/orchestrator/engine.py b/src/orchestrator/engine.py
index 31f8ed8..9018808 100644
--- a/src/orchestrator/engine.py
+++ b/src/orchestrator/engine.py
@@ -123,6 +123,23 @@ class OrchestratorEngine:
             # Direct response — no plan needed (saludo, pregunta simple)
             if isinstance(plan_result, str):
                 logger.info("Planner returned direct response for task %s", task.task_id)
+
+                # Save to task history so conversation context is preserved
+                session.task_history.append({
+                    "task_id": task.task_id,
+                    "objective": message,
+                    "status": "completed",
+                    "steps": 0,
+                    "facts": [],
+                    "key_data": {},
+                    "tools_used": [],
+                    "artifacts_count": 0,
+                    "summary": f"User: {message[:200]} → Agent: {plan_result[:200]}",
+                    "review": "",
+                })
+                if len(session.task_history) > 20:
+                    session.task_history = session.task_history[-20:]
+
                 task.status = TaskStatus.COMPLETED
                 session.complete_task()