universal-mcp-agents 0.1.9__py3-none-any.whl → 0.1.11__py3-none-any.whl

universal_mcp/agents/planner/__init__.py

@@ -26,9 +26,7 @@ class PlannerAgent(BaseAgent):
         self.executor_agent_cls = executor_agent_cls
 
     def _build_system_message(self):
-        return DEVELOPER_PROMPT.format(
-            name=self.name, instructions=self.instructions
-        )
+        return DEVELOPER_PROMPT.format(name=self.name, instructions=self.instructions)
 
     async def _build_graph(self):
         return build_graph(

universal_mcp/agents/planner/graph.py

@@ -18,9 +18,7 @@ def build_graph(llm, registry, instructions, model, executor_agent_cls):
         task = state["messages"][-1].content
         logger.info(f"Running tool finder for task: {task}")
         tool_finder_graph = build_tool_node_graph(llm, registry)
-        tool_finder_state = await tool_finder_graph.ainvoke(
-            {"original_task": task, "messages": state["messages"]}
-        )
+        tool_finder_state = await tool_finder_graph.ainvoke({"original_task": task, "messages": state["messages"]})
 
         if not tool_finder_state.get("apps_required"):
             logger.info("Tool finder determined no apps are required.")

universal_mcp/agents/react.py

@@ -4,6 +4,7 @@ from loguru import logger
 from universal_mcp.agentr.registry import AgentrRegistry
 from universal_mcp.tools.registry import ToolRegistry
 from universal_mcp.types import ToolConfig, ToolFormat
+from rich import print
 
 from universal_mcp.agents.base import BaseAgent
 from universal_mcp.agents.llm import load_chat_model
@@ -40,7 +41,18 @@ class ReactAgent(BaseAgent):
         self.llm = load_chat_model(model)
         self.tools = tools or {}
         if "ui" not in self.tools:
-            self.tools["ui"] = ["create_bar_chart", "create_line_chart", "create_pie_chart", "create_table", "http_get", "http_post", "http_put", "http_delete", "http_patch", "read_file"]
+            self.tools["ui"] = [
+                "create_bar_chart",
+                "create_line_chart",
+                "create_pie_chart",
+                "create_table",
+                "http_get",
+                "http_post",
+                "http_put",
+                "http_delete",
+                "http_patch",
+                "read_file",
+            ]
         self.max_iterations = max_iterations
         self.registry = registry
 
@@ -54,7 +66,6 @@ class ReactAgent(BaseAgent):
         else:
             tools = []
 
-
         logger.debug(f"Initialized ReactAgent: name={self.name}, model={self.model}")
         return create_react_agent(
             self.llm,
@@ -75,10 +86,7 @@ async def main():
         tools={"google-mail": ["send_email"]},
         registry=AgentrRegistry(),
     )
-    result = await agent.invoke(
-        "Send an email with the subject 'testing react agent' to manoj@agentr.dev"
-    )
-    from rich import print
+    result = await agent.invoke("Send an email with the subject 'testing react agent' to manoj@agentr.dev")
 
     print(messages_to_list(result["messages"]))
 

universal_mcp/agents/shared/prompts.py

@@ -6,7 +6,7 @@ You are an expert planner. Your goal is to consolidate a complex user request in
 2.  **Focus on Data Handoffs:** A good decomposition often involves one sub-task to *retrieve* information and a subsequent sub-task to *use* that information.
 3.  **Assume Internal Capabilities:** Do NOT create sub-tasks for abstract cognitive work like 'summarize' or 'analyze'.
 4.  **Simplify Single Actions:** If the user's task is already a single, simple action, the output should be a single sub-task that concisely describes that action. Do not make it identical to the user's input.
-5.  **General purpose sub tasks:** You also need to realise that these subtasks are going to be used to search for tools and apps. And the names and description of these tools and apps are going to be general in nature so the sub tasks should not be too specific. The task which you will get may be specific in nature but the sub taks must be general. 
+5.  **General purpose sub tasks:** You also need to realise that these subtasks are going to be used to search for tools and apps. And the names and description of these tools and apps are going to be general in nature so the sub tasks should not be too specific. The task which you will get may be specific in nature but the sub taks must be general.
 **--- EXAMPLES ---**
 
 **EXAMPLE 1:**
@@ -116,7 +116,7 @@ You are an AI assistant that selects the most appropriate tool(s) from a list to
 **INSTRUCTIONS:**
 1.  Carefully review the sub-task to understand the required action.
 2.  Examine the list of available tools and their descriptions.
-3.  Select the best tool ID that matches the sub-task. You are encouraged to select multiple tools if there are multiple tools with similar capabilties 
+3.  Select the best tool ID that matches the sub-task. You are encouraged to select multiple tools if there are multiple tools with similar capabilties
 or names. It is always good to have more tools than having insufficent tools.
 4.  If no tool is a good fit, return an empty list.
 5.  Only return the tool IDs.
@@ -129,4 +129,4 @@ or names. It is always good to have more tools than having insufficent tools.
 {tool_candidates}
 
 **YOUR SELECTED TOOL ID(s):**
-"""
+"""

universal_mcp/agents/shared/tool_node.py

@@ -1,5 +1,5 @@
 import asyncio
-from typing import Annotated, TypedDict, List, Dict
+from typing import Annotated, TypedDict
 
 from langchain_core.language_models import BaseChatModel
 from langchain_core.messages import AIMessage, AnyMessage, HumanMessage
@@ -17,43 +17,50 @@ from universal_mcp.agents.shared.prompts import (
     TOOL_SELECTION_PROMPT,
 )
 
-
 MAX_DECOMPOSITION_ATTEMPTS = 2
 
 # --- Pydantic Models for Structured LLM Outputs ---
 
+
 class TaskDecomposition(BaseModel):
-    sub_tasks: List[str] = Field(description="A list of sub-task descriptions.")
+    sub_tasks: list[str] = Field(description="A list of sub-task descriptions.")
+
 
 class SearchQuery(BaseModel):
     query: str = Field(description="A concise search query.")
 
+
 class ToolSelection(BaseModel):
-    tool_ids: List[str] = Field(description="The IDs of the selected tools.")
+    tool_ids: list[str] = Field(description="The IDs of the selected tools.")
 
 
 # --- LangGraph Agent State ---
 
+
 class SubTask(TypedDict, total=False):
     """Represents a single step in the execution plan."""
+
     task: str
     status: str  # "pending", "success", "failed"
     app_id: str
-    tool_ids: List[str]
+    tool_ids: list[str]
     reasoning: str
 
+
 class AgentState(TypedDict):
     """The central state of our agent graph."""
+
     original_task: str
     decomposition_attempts: int
     failed_sub_task_info: str  # To inform re-decomposition
-    sub_tasks: List[SubTask]
-    execution_plan: List[SubTask]
+    sub_tasks: list[SubTask]
+    execution_plan: list[SubTask]
     messages: Annotated[list[AnyMessage], add_messages]
 
 
 # --- Graph Builder ---
 
+
 def build_tool_node_graph(llm: BaseChatModel, registry: ToolRegistry) -> StateGraph:
     """Builds the adaptive LangGraph workflow for tool selection."""
 
@@ -65,19 +72,14 @@ def build_tool_node_graph(llm: BaseChatModel, registry: ToolRegistry) -> StateGr
 
         if attempts > 0 and failed_info:
             logger.warning(f"Revising decomposition. Attempt {attempts + 1}.")
-            prompt = REVISE_DECOMPOSITION_PROMPT.format(
-                task=task, failed_sub_task=failed_info
-            )
+            prompt = REVISE_DECOMPOSITION_PROMPT.format(task=task, failed_sub_task=failed_info)
         else:
             logger.info("Performing initial task decomposition.")
             prompt = TASK_DECOMPOSITION_PROMPT.format(task=task)
 
         response = await llm.with_structured_output(TaskDecomposition).ainvoke(prompt)
-        sub_tasks = [
-            {"task": sub_task_str, "status": "pending"}
-            for sub_task_str in response.sub_tasks
-        ]
-        
+        sub_tasks = [{"task": sub_task_str, "status": "pending"} for sub_task_str in response.sub_tasks]
+
         return {
             "sub_tasks": sub_tasks,
             "decomposition_attempts": attempts + 1,
@@ -99,16 +101,13 @@ def build_tool_node_graph(llm: BaseChatModel, registry: ToolRegistry) -> StateGr
                 plan_context_str = "None. This is the first step."
             else:
                 context_lines = [
-                    f"- The sub-task '{step['task']}' was assigned to app '{step['app_id']}'."
-                    for step in current_plan
+                    f"- The sub-task '{step['task']}' was assigned to app '{step['app_id']}'." for step in current_plan
                 ]
                 plan_context_str = "\n".join(context_lines)
 
             # 2. Generate the App-specific query using the NEW full-context prompt
             app_query_prompt = APP_SEARCH_QUERY_PROMPT.format(
-                original_task=original_task,
-                plan_context=plan_context_str,
-                sub_task=task_desc
+                original_task=original_task, plan_context=plan_context_str, sub_task=task_desc
             )
             app_query_response = await llm.with_structured_output(SearchQuery).ainvoke(app_query_prompt)
             app_search_query = app_query_response.query
@@ -142,12 +141,14 @@ def build_tool_node_graph(llm: BaseChatModel, registry: ToolRegistry) -> StateGr
 
                 if selection_response.tool_ids:
                     logger.success(f"Found and selected tool(s) {selection_response.tool_ids} in app '{app_id}'.")
-                    sub_task.update({
-                        "status": "success",
-                        "app_id": app_id,
-                        "tool_ids": selection_response.tool_ids,
-                        "reasoning": f"Selected tool(s) {selection_response.tool_ids} from app '{app_id}' for sub-task."
-                    })
+                    sub_task.update(
+                        {
+                            "status": "success",
+                            "app_id": app_id,
+                            "tool_ids": selection_response.tool_ids,
+                            "reasoning": f"Selected tool(s) {selection_response.tool_ids} from app '{app_id}' for sub-task.",
+                        }
+                    )
                     current_plan.append(sub_task)
                     tool_found = True
                     break
@@ -157,12 +158,16 @@ def build_tool_node_graph(llm: BaseChatModel, registry: ToolRegistry) -> StateGr
                 return {"failed_sub_task_info": task_desc, "sub_tasks": []}
 
         return {"execution_plan": current_plan, "sub_tasks": []}
-    
+
     def _handle_planning_failure(state: AgentState) -> AgentState:
         """Handles the case where all decomposition attempts have failed."""
         logger.error("Maximum decomposition attempts reached. Planning failed.")
         return {
-            "messages": [AIMessage(content="I am unable to create a complete plan for this task with the available tools. Please try rephrasing your request.")]
+            "messages": [
+                AIMessage(
+                    content="I am unable to create a complete plan for this task with the available tools. Please try rephrasing your request."
+                )
+            ]
         }
 
     def _consolidate_plan(state: AgentState) -> AgentState:
@@ -172,7 +177,7 @@ def build_tool_node_graph(llm: BaseChatModel, registry: ToolRegistry) -> StateGr
         """
         logger.info("Consolidating final execution plan.")
         plan = state["execution_plan"]
-        merged_apps: Dict[str, SubTask] = {}
+        merged_apps: dict[str, SubTask] = {}
 
         for step in plan:
             app_id = step["app_id"]
@@ -189,9 +194,8 @@ def build_tool_node_graph(llm: BaseChatModel, registry: ToolRegistry) -> StateGr
         for app_id, step_data in merged_apps.items():
             step_data["tool_ids"] = sorted(list(step_data["tool_ids"]))
             final_plan.append(step_data)
-        
-        return {"execution_plan": final_plan}
 
+        return {"execution_plan": final_plan}
 
     # --- Graph Definition ---
 
@@ -199,26 +203,26 @@ def build_tool_node_graph(llm: BaseChatModel, registry: ToolRegistry) -> StateGr
 
     workflow.add_node("decompose_task", _decompose_task)
     workflow.add_node("resolve_sub_tasks", _resolve_sub_tasks)
-    workflow.add_node("consolidate_plan", _consolidate_plan) # NEW NODE
+    workflow.add_node("consolidate_plan", _consolidate_plan)  # NEW NODE
     workflow.add_node("handle_planning_failure", _handle_planning_failure)
 
     workflow.set_entry_point("decompose_task")
 
     def should_continue(state: AgentState):
-        if not state.get("sub_tasks"): # Resolution failed or succeeded
+        if not state.get("sub_tasks"):  # Resolution failed or succeeded
             if state.get("execution_plan"):
-                return "consolidate_plan" # MODIFIED: Go to consolidate on success
+                return "consolidate_plan"  # MODIFIED: Go to consolidate on success
             elif state["decomposition_attempts"] >= MAX_DECOMPOSITION_ATTEMPTS:
                 return "handle_planning_failure"
             else:
-                return "decompose_task" # Re-try decomposition
+                return "decompose_task"  # Re-try decomposition
         else:
             return "resolve_sub_tasks"
 
     workflow.add_conditional_edges("decompose_task", lambda s: "resolve_sub_tasks")
     workflow.add_conditional_edges("resolve_sub_tasks", should_continue)
-    
-    workflow.add_edge("consolidate_plan", END) # NEW EDGE
+
+    workflow.add_edge("consolidate_plan", END)  # NEW EDGE
     workflow.add_edge("handle_planning_failure", END)
 
     return workflow.compile()
@@ -227,15 +231,16 @@ def build_tool_node_graph(llm: BaseChatModel, registry: ToolRegistry) -> StateGr
 async def main():
     """Main function to run the agent."""
     from universal_mcp.agentr.registry import AgentrRegistry
+
     from universal_mcp.agents.llm import load_chat_model
 
     registry = AgentrRegistry()
     llm = load_chat_model("anthropic/claude-4-sonnet-20250514")
-    
+
     graph = build_tool_node_graph(llm, registry)
 
     task = "Find my latest order confirmation in Gmail, search for reviews of the main product on perplexity, and then send an email to ankit@agentr.dev telling about the reviews"
-    
+
     initial_state = {
         "original_task": task,
         "messages": [HumanMessage(content=task)],
@@ -244,17 +249,12 @@ async def main():
 
     final_state = await graph.ainvoke(initial_state)
 
-    print("\n--- Final Agent State ---")
     if final_state.get("execution_plan"):
-        print("Successfully created a consolidated execution plan:")
         for step in final_state["execution_plan"]:
-            print(f"- Sub-task: {step['task']}")
-            print(f"  - App: {step['app_id']}")
-            print(f"  - Tool(s): {', '.join(step['tool_ids'])}")
+            pass
     else:
-        print("Failed to create an execution plan.")
-        print(f"Final message: {final_state['messages'][-1].content}")
+        pass
 
 
 if __name__ == "__main__":
-    asyncio.run(main())
+    asyncio.run(main())

universal_mcp/agents/simple.py

@@ -4,6 +4,7 @@ from typing import Annotated
 from langgraph.checkpoint.base import BaseCheckpointSaver
 from langgraph.graph import END, START, StateGraph
 from langgraph.graph.message import add_messages
+from rich import print
 from typing_extensions import TypedDict
 
 from universal_mcp.agents.base import BaseAgent
@@ -52,10 +53,10 @@ class SimpleAgent(BaseAgent):
         graph_builder.add_edge("chatbot", END)
         return graph_builder.compile(checkpointer=self.memory)
 
+
 async def main():
     agent = SimpleAgent("Simple Agent", "Act as a 14 year old kid, reply in Gen-Z lingo", "azure/gpt-5-mini")
     output = await agent.invoke("What is the capital of France?")
-    from rich import print
     print(messages_to_list(output["messages"]))
 
 

universal_mcp/agents/utils.py

@@ -8,9 +8,6 @@ from rich.markdown import Markdown
 from rich.panel import Panel
 from rich.prompt import Prompt
 from rich.table import Table
-from universal_mcp.tools.manager import ToolManager
-from universal_mcp.types import ToolFormat
-
 
 
 class RichCLI:
@@ -28,9 +25,7 @@ Available commands:
 - `/tools` - List available tools
 - `/exit` - Exit the application
         """
-        self.console.print(
-            Panel(Markdown(welcome_text), title="🤖 AI Agent CLI", border_style="blue")
-        )
+        self.console.print(Panel(Markdown(welcome_text), title="🤖 AI Agent CLI", border_style="blue"))
 
     def display_agent_response(self, response: str, agent_name: str):
         """Display agent response with formatting"""
@@ -54,13 +49,9 @@ Available commands:
                     # Check if type has changed and reset content if so
                     if self.type_ != type_:
                         if type_ == "thinking":
-                            self.content += (
-                                "\n[bold yellow]💭 Thinking:[/bold yellow] :"
-                            )
+                            self.content += "\n[bold yellow]💭 Thinking:[/bold yellow] :"
                         elif type_ == "text":
-                            self.content += (
-                                f"\n[bold green]🤖 {agent_name}[/bold green] :"
-                            )
+                            self.content += f"\n[bold green]🤖 {agent_name}[/bold green] :"
                         self.type_ = type_
                     self.content += chunk
                     content_text = "".join(self.content)
@@ -120,9 +111,7 @@ Available commands:
             value = Prompt.ask(interrupt.value["question"])
             return value
         elif interrupt_type == "bool":
-            value = Prompt.ask(
-                interrupt.value["question"], choices=["y", "n"], default="y"
-            )
+            value = Prompt.ask(interrupt.value["question"], choices=["y", "n"], default="y")
             return value
         elif interrupt_type == "choice":
             value = Prompt.ask(

universal_mcp/applications/ui/app.py

@@ -134,9 +134,7 @@ class UiApp(BaseApplication):
                 "headers": dict(response.headers),
             }
 
-    def http_get(
-        self, url: str, headers: dict | None = None, query_params: dict | None = None
-    ):
+    def http_get(self, url: str, headers: dict | None = None, query_params: dict | None = None):
         """
         Perform a GET request to the specified URL with optional parameters.
 
@@ -150,16 +148,12 @@ class UiApp(BaseApplication):
         Tags:
             get, important
         """
-        logger.debug(
-            f"GET request to {url} with headers {headers} and query params {query_params}"
-        )
+        logger.debug(f"GET request to {url} with headers {headers} and query params {query_params}")
         response = httpx.get(url, params=query_params, headers=headers)
         response.raise_for_status()
         return self._handle_response(response)
 
-    def http_post(
-        self, url: str, headers: dict | None = None, body: dict | None = None
-    ):
+    def http_post(self, url: str, headers: dict | None = None, body: dict | None = None):
         """
         Perform a POST request to the specified URL with optional parameters.
 
@@ -197,9 +191,7 @@ class UiApp(BaseApplication):
         response.raise_for_status()
         return self._handle_response(response)
 
-    def http_delete(
-        self, url: str, headers: dict | None = None, body: dict | None = None
-    ):
+    def http_delete(self, url: str, headers: dict | None = None, body: dict | None = None):
         """
         Perform a DELETE request to the specified URL with optional parameters.
 
@@ -218,9 +210,7 @@ class UiApp(BaseApplication):
         response.raise_for_status()
         return self._handle_response(response)
 
-    def http_patch(
-        self, url: str, headers: dict | None = None, body: dict | None = None
-    ):
+    def http_patch(self, url: str, headers: dict | None = None, body: dict | None = None):
         """
         Perform a PATCH request to the specified URL with optional parameters.
 

{universal_mcp_agents-0.1.9.dist-info → universal_mcp_agents-0.1.11.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: universal-mcp-agents
-Version: 0.1.9
+Version: 0.1.11
 Summary: Add your description here
 Project-URL: Homepage, https://github.com/universal-mcp/applications
 Project-URL: Repository, https://github.com/universal-mcp/applications
@@ -11,6 +11,7 @@ Requires-Dist: langchain-anthropic>=0.3.19
 Requires-Dist: langchain-google-genai>=2.1.10
 Requires-Dist: langchain-openai>=0.3.32
 Requires-Dist: langgraph>=0.6.6
+Requires-Dist: typer>=0.17.4
 Requires-Dist: universal-mcp-applications>=0.1.14
 Requires-Dist: universal-mcp>=0.1.24rc21
 Provides-Extra: dev

universal_mcp_agents-0.1.11.dist-info/RECORD

@@ -0,0 +1,42 @@
+universal_mcp/agents/__init__.py,sha256=oPoHMITGbHN4Ey68ZrVbmy7sNzVYhoXQgVII9fZwqL8,1245
+universal_mcp/agents/base.py,sha256=KXBxf3TXrVHi-wBVD-cs6PSKfMtUnm73l-hC83FjOog,6753
+universal_mcp/agents/builder.py,sha256=Xl_dGmzbtop3lICH2njnN6yxFF0SnEGY8u1tOIJy2Pk,8677
+universal_mcp/agents/cli.py,sha256=-luC55FHCTwnpcRgqnV95yQa-mGlLYDlseUjLqFrAfs,1014
+universal_mcp/agents/hil.py,sha256=_xLlBte4v5ex-RxXy5H3LqwFhtd3KE1QUHX1QDGIl2w,3760
+universal_mcp/agents/llm.py,sha256=hVRwjZs3MHl5_3BWedmurs2Jt1oZDfFX0Zj9F8KH7fk,1787
+universal_mcp/agents/react.py,sha256=0OZvdBTnQdFF3Wliv2l4wiF2BMd0VG1cVflOaWDC8r0,3166
+universal_mcp/agents/simple.py,sha256=W5_zb2DAGtAx_p5weSQGIC3gheZwxhmwZBuKKBTjd04,1938
+universal_mcp/agents/utils.py,sha256=lnRmP6HBIme4W-iKy1BOb9gTyb9-ooJxXwCyWAQFP9k,4712
+universal_mcp/agents/bigtool2/__init__.py,sha256=i4virR9r1_1FcS_-iuSHZWgEzYZwOroT6J44qPb0ZgM,2462
+universal_mcp/agents/bigtool2/__main__.py,sha256=t6fWhLh3SnpN_05cww3LA_r_5Rb0gaF_U4FH1Mpsv1Y,655
+universal_mcp/agents/bigtool2/agent.py,sha256=4GIQIy2VQgdXOezmET8G7tvP_37Vv8C027bGdGXJbTI,437
+universal_mcp/agents/bigtool2/graph.py,sha256=FuW1XyTIr1aUbAC4ea8JWRR0JENvLfZGUSgQFmk0h3A,6544
+universal_mcp/agents/bigtool2/meta_tools.py,sha256=02xOsGdxZpXBirn2KWk63UqYPQjI41nQ2KGj2zKBf7Y,5306
+universal_mcp/agents/bigtool2/prompts.py,sha256=rQFtZDkwU9z8d4PWdt6jpohGhyab658Xvk8hvNVBFBA,1843
+universal_mcp/agents/bigtool2/state.py,sha256=TQeGZD99okclkoCh5oz-VYIlEsC9yLQyDpnBnm7QCN8,759
+universal_mcp/agents/bigtoolcache/__init__.py,sha256=qrUwYqhZNE0PdmUCGc5jRQt45Tr1xPDSfBxCPJA1DDM,2260
+universal_mcp/agents/bigtoolcache/__main__.py,sha256=noqT7Nqr17q7Eeiib-Dk2qG6AH4LkNyveeU_ceNDfMA,473
+universal_mcp/agents/bigtoolcache/agent.py,sha256=NCF6fdBPwVcaOyhv8xbAy5DBqbGxJiv8_fLqKib7QJc,267
+universal_mcp/agents/bigtoolcache/context.py,sha256=ny7gd-vvVpUOYAeQbAEUT0A6Vm6Nn2qGywxTzPBzYFg,929
+universal_mcp/agents/bigtoolcache/graph.py,sha256=9KUjsVu5dtPdnyNkqJ5LsMnfH-IF_4H2me_L5aHgsug,4315
+universal_mcp/agents/bigtoolcache/prompts.py,sha256=Rz30qNGdscDG65vMj9d0Vfe7X1pQjBDQBBNc3BuyC94,1886
+universal_mcp/agents/bigtoolcache/state.py,sha256=TQeGZD99okclkoCh5oz-VYIlEsC9yLQyDpnBnm7QCN8,759
+universal_mcp/agents/bigtoolcache/tools.py,sha256=ynyEj9mVwKKDhxm76sjspyH51SFi63g2Vydi39pY0qY,5562
+universal_mcp/agents/codeact/__init__.py,sha256=rLE8gvOo5H4YSr71DRq76b3RV3uuotxuAy_VnBVaVwk,60
+universal_mcp/agents/codeact/__main__.py,sha256=FRfIkgcZfawP-M66v4ePijA6J2fs7nQv92G_8cj5qYA,1142
+universal_mcp/agents/codeact/agent.py,sha256=yB99aTRHWOdl4b67UxRstuV7WplrTjUZuqP1odwSS5o,5586
+universal_mcp/agents/codeact/prompts.py,sha256=6NZkWQCaS7X7CLOTV-hMZgeWa1jf6iKbAT00INuluM0,4668
+universal_mcp/agents/codeact/sandbox.py,sha256=pG6M1elzWw4KS46ewl-2d_U9Ap8HnHJDY9Rna_76sRI,1489
+universal_mcp/agents/codeact/state.py,sha256=K25HcEljNNJDaeUgG5P68Kj752XzWTh8BwILY8w_EtE,357
+universal_mcp/agents/codeact/utils.py,sha256=JUbT_HYGS_D1BzmzoVpORIe7SGur1KgJguTZ_1tZ4JY,1918
+universal_mcp/agents/planner/__init__.py,sha256=9P1UL-ABvrTIWTJ8wcvZmkqT8uyROZxsmUFhpjTK-Q4,1313
+universal_mcp/agents/planner/__main__.py,sha256=OfhTfYDZK_ZUfc8sX-Sa6TWk-dNqD2rl13Ln64mNAtw,771
+universal_mcp/agents/planner/graph.py,sha256=70hhIoEZOcYojpiyVSCedgYpnmxVP7aqdn8s6VBu-D4,3228
+universal_mcp/agents/planner/prompts.py,sha256=_JoHqiAvswtqCDu90AGUHmfsu8eWE1-_yI4LLn3pqMU,657
+universal_mcp/agents/planner/state.py,sha256=qqyp-jSGsCxe1US-PRLT4-y1sITAcVE6nCMlQLnvop0,278
+universal_mcp/agents/shared/prompts.py,sha256=VOsXSUEwBXPaAuxJTUF6bgDGr41u6uctUNQSMRt_OJc,6414
+universal_mcp/agents/shared/tool_node.py,sha256=Ua_wzMt4YgIx4zLp3_ZCow-28qORwrZ2FvKqLPt3RlI,10415
+universal_mcp/applications/ui/app.py,sha256=uaS1KrwrGxw9oexdLj2Jok77DrZQAmby3uVxCONQyV8,11276
+universal_mcp_agents-0.1.11.dist-info/METADATA,sha256=MosDAsvQTSvUr41527xNHF8QpEeatOh7g0bpBDr7lW0,878
+universal_mcp_agents-0.1.11.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
+universal_mcp_agents-0.1.11.dist-info/RECORD,,

universal_mcp/agents/autoagent/__init__.py

@@ -1,30 +0,0 @@
-from langgraph.checkpoint.base import BaseCheckpointSaver
-from universal_mcp.tools.registry import ToolRegistry
-
-from universal_mcp.agents.autoagent.graph import build_graph
-from universal_mcp.agents.base import BaseAgent
-
-
-class AutoAgent(BaseAgent):
-    def __init__(
-        self,
-        name: str,
-        instructions: str,
-        model: str,
-        memory: BaseCheckpointSaver | None = None,
-        registry: ToolRegistry | None = None,
-        **kwargs,
-    ):
-        super().__init__(name, instructions, model, memory, **kwargs)
-        self.tool_registry = registry
-
-    async def _build_graph(self):
-        builder = await build_graph(self.tool_registry, self.instructions)
-        return builder.compile(checkpointer=self.memory)
-
-    @property
-    def graph(self):
-        return self._graph
-
-
-__all__ = ["AutoAgent"]

universal_mcp/agents/autoagent/__main__.py

@@ -1,25 +0,0 @@
-import asyncio
-
-from loguru import logger
-from universal_mcp.agentr.registry import AgentrRegistry
-
-from universal_mcp.agents.autoagent import AutoAgent
-
-
-async def main():
-    agent = AutoAgent(
-        name="autoagent",
-        instructions="You are a helpful assistant that can use tools to help the user.",
-        model="azure/gpt-4.1",
-        registry=AgentrRegistry(),
-    )
-    async for event in agent.stream(
-        user_input="Send an email to manoj@agentr.dev",
-        thread_id="test123",
-    ):
-        logger.info(event.content)
-    # from loguru import logger; logger.debug(result)
-
-
-if __name__ == "__main__":
-    asyncio.run(main())

universal_mcp/agents/autoagent/context.py

@@ -1,26 +0,0 @@
-from dataclasses import dataclass, field
-from typing import Annotated
-
-from universal_mcp.agents.autoagent.prompts import SYSTEM_PROMPT
-
-
-@dataclass(kw_only=True)
-class Context:
-    """The context for the agent."""
-
-    system_prompt: str = field(
-        default=SYSTEM_PROMPT,
-        metadata={
-            "description": "The system prompt to use for the agent's interactions. "
-            "This prompt sets the context and behavior for the agent."
-        },
-    )
-
-    model: Annotated[str, {"__template_metadata__": {"kind": "llm"}}] = field(
-        default="anthropic/claude-4-sonnet-20250514",
-        # default="vertex/gemini-2.5-flash",
-        metadata={
-            "description": "The name of the language model to use for the agent's main interactions. "
-            "Should be in the form: provider/model-name."
-        },
-    )