universal-mcp-agents 0.1.19rc1__py3-none-any.whl → 0.1.24rc3__py3-none-any.whl

universal_mcp/agents/codeact0/agent.py

@@ -1,37 +1,46 @@
-import inspect
-from collections.abc import Callable
+import copy
+import json
+import re
+import uuid
 from typing import Literal, cast
 
-from langchain_core.messages import AIMessage, ToolMessage
-from langchain_core.tools import StructuredTool
-from langchain_core.tools import tool as create_tool
+from langchain_anthropic import ChatAnthropic
+from langchain_core.messages import AIMessage, HumanMessage, ToolMessage
 from langgraph.checkpoint.base import BaseCheckpointSaver
 from langgraph.graph import START, StateGraph
-from langgraph.types import Command, RetryPolicy
+from langgraph.types import Command, RetryPolicy, StreamWriter
 from universal_mcp.tools.registry import ToolRegistry
-from universal_mcp.types import ToolConfig, ToolFormat
+from universal_mcp.types import ToolFormat
 
 from universal_mcp.agents.base import BaseAgent
-from universal_mcp.agents.codeact0.llm_tool import ai_classify, call_llm, data_extractor, smart_print
+from universal_mcp.agents.codeact0.llm_tool import smart_print
 from universal_mcp.agents.codeact0.prompts import (
+    AGENT_BUILDER_GENERATING_PROMPT,
+    AGENT_BUILDER_META_PROMPT,
+    AGENT_BUILDER_PLANNING_PROMPT,
+    build_tool_definitions,
     create_default_prompt,
 )
-from universal_mcp.agents.codeact0.sandbox import eval_unsafe, execute_ipython_cell
-from universal_mcp.agents.codeact0.state import CodeActState
-from universal_mcp.agents.codeact0.utils import inject_context, smart_truncate
+from universal_mcp.agents.codeact0.sandbox import eval_unsafe, execute_ipython_cell, handle_execute_ipython_cell
+from universal_mcp.agents.codeact0.state import AgentBuilderCode, AgentBuilderMeta, AgentBuilderPlan, CodeActState
+from universal_mcp.agents.codeact0.tools import (
+    create_meta_tools,
+    enter_agent_builder_mode,
+)
+from universal_mcp.agents.codeact0.utils import build_anthropic_cache_message, get_connected_apps_string, strip_thinking, extract_plan_parameters
 from universal_mcp.agents.llm import load_chat_model
-from universal_mcp.agents.utils import filter_retry_on
+from universal_mcp.agents.utils import convert_tool_ids_to_dict, filter_retry_on, get_message_text
 
 
-class CodeActAgent(BaseAgent):
+class CodeActPlaybookAgent(BaseAgent):
     def __init__(
         self,
         name: str,
         instructions: str,
         model: str,
         memory: BaseCheckpointSaver | None = None,
-        tools: ToolConfig | None = None,
         registry: ToolRegistry | None = None,
+        agent_builder_registry: object | None = None,
         sandbox_timeout: int = 20,
         **kwargs,
     ):
@@ -42,103 +51,442 @@ class CodeActAgent(BaseAgent):
             memory=memory,
             **kwargs,
         )
-        self.model_instance = load_chat_model(model, thinking=True)
-        self.tools_config = tools or {}
+        self.model_instance = load_chat_model(model)
+        self.agent_builder_model_instance = load_chat_model("anthropic:claude-sonnet-4-5-20250929", thinking=False)
         self.registry = registry
+        self.agent_builder_registry = agent_builder_registry
+        self.agent = agent_builder_registry.get_agent() if agent_builder_registry else None
+
+        self.tools_config = self.agent.tools if self.agent else {}
         self.eval_fn = eval_unsafe
         self.sandbox_timeout = sandbox_timeout
-        self.processed_tools: list[StructuredTool | Callable] = []
-
-        # TODO(manoj): Use toolformat native instead of langchain
-        # TODO(manoj, later): Add better sandboxing
-        # Old Nishant TODO s:
-        # - Make codeact faster by calling upto API call (this done but should be tested)
-        # - Add support for async eval_fn
-        # - Throw Error if code snippet is too long (> 1000 characters) and suggest to split it into smaller parts
-        # - Multiple models from config
-
-    async def _build_graph(self):
-        exported_tools = []
+        self.default_tools_config = {
+            "llm": ["generate_text", "classify_data", "extract_data", "call_llm"],
+        }
+        self.final_instructions = ""
+        self.tools_context = {}
+        self.eval_mode = kwargs.get("eval_mode", False)
+
+    async def _build_graph(self):  # noqa: PLR0915
+        """Build the graph for the CodeAct Playbook Agent."""
+        meta_tools = create_meta_tools(self.registry)
+        self.additional_tools = [
+            smart_print,
+            meta_tools["web_search"],
+            meta_tools["read_file"],
+            meta_tools["save_file"],
+            meta_tools["upload_file"],
+        ]
+
         if self.tools_config:
-            if not self.registry:
-                raise ValueError("Tools are configured but no registry is provided")
-            # Langchain tools are fine
-            exported_tools = await self.registry.export_tools(self.tools_config, ToolFormat.LANGCHAIN)
-        additional_tools = [smart_print, data_extractor, ai_classify, call_llm]
-        additional_tools = [t if isinstance(t, StructuredTool) else create_tool(t) for t in additional_tools]
-        self.instructions, self.tools_context = create_default_prompt(
-            exported_tools, additional_tools, self.instructions
-        )
+            await self.registry.load_tools(self.tools_config)  # Load provided tools
+        if self.default_tools_config:
+            await self.registry.load_tools(self.default_tools_config)  # Load default tools
 
-        def call_model(state: CodeActState) -> Command[Literal["sandbox"]]:
-            messages = [{"role": "system", "content": self.instructions}] + state["messages"]
+        async def call_model(state: CodeActState) -> Command[Literal["execute_tools"]]:
+            """This node now only ever binds the four meta-tools to the LLM."""
+            messages = build_anthropic_cache_message(self.final_instructions) + state["messages"]
 
-            # Run the model and potentially loop for reflection
-            model_with_tools = self.model_instance.bind_tools(tools=[execute_ipython_cell], tool_choice="auto")
-            response = cast(AIMessage, model_with_tools.invoke(messages))
+            agent_facing_tools = [
+                execute_ipython_cell,
+                enter_agent_builder_mode,
+                meta_tools["search_functions"],
+                meta_tools["load_functions"],
+            ]
 
+            if isinstance(self.model_instance, ChatAnthropic):
+                model_with_tools = self.model_instance.bind_tools(
+                    tools=agent_facing_tools,
+                    tool_choice="auto",
+                    cache_control={"type": "ephemeral", "ttl": "1h"},
+                )
+                if isinstance(messages[-1].content, str):
+                    pass
+                else:
+                    last = copy.deepcopy(messages[-1])
+                    last.content[-1]["cache_control"] = {"type": "ephemeral", "ttl": "5m"}
+                    messages[-1] = last
+            else:
+                model_with_tools = self.model_instance.bind_tools(
+                    tools=agent_facing_tools,
+                    tool_choice="auto",
+                )
+            response = cast(AIMessage, await model_with_tools.ainvoke(messages))
             if response.tool_calls:
-                if len(response.tool_calls) > 1:
-                    raise Exception("Not possible in Claude with llm.bind_tools(tools=tools, tool_choice='auto')")
-                if response.tool_calls[0]["name"] != "execute_ipython_cell":
-                    raise Exception(
-                        f"Unexpected tool call: {response.tool_calls[0]['name']}. Expected 'execute_ipython_cell'."
-                    )
-                if (
-                    response.tool_calls[0]["args"].get("snippet") is None
-                    or not response.tool_calls[0]["args"]["snippet"].strip()
-                ):
-                    raise Exception("Tool call 'execute_ipython_cell' requires a non-empty 'snippet' argument.")
-                return Command(goto="sandbox", update={"messages": [response]})
+                return Command(goto="execute_tools", update={"messages": [response]})
             else:
-                return Command(update={"messages": [response]})
-
-        # If eval_fn is a async, we define async node function.
-        if inspect.iscoroutinefunction(self.eval_fn):
-            raise ValueError("eval_fn must be a synchronous function, not a coroutine.")
-            # async def sandbox(state: StateSchema):
-            #     existing_context = state.get("context", {})
-            #     context = {**existing_context, **tools_context}
-            #     # Execute the script in the sandbox
-            #     output, new_vars = await eval_fn(state["script"], context)
-            #     new_context = {**existing_context, **new_vars}
-            #     return {
-            #         "messages": [{"role": "user", "content": output}],
-            #         "context": new_context,
-            #     }
-        else:
-
-            def sandbox(state: CodeActState) -> Command[Literal["call_model"]]:
-                tool_call = state["messages"][-1].tool_calls[0]  # type: ignore
-                code = tool_call["args"]["snippet"]
-                previous_add_context = state.get("add_context", {})
-                add_context = inject_context(previous_add_context, self.tools_context)
-                existing_context = state.get("context", {})
-                context = {**existing_context, **add_context}
-                # Execute the script in the sandbox
-
-                output, new_context, new_add_context = self.eval_fn(
-                    code, context, previous_add_context, 180
-                )  # default timeout 3 min
-                output = smart_truncate(output)
+                return Command(update={"messages": [response], "model_with_tools": model_with_tools})
+
+        async def execute_tools(state: CodeActState) -> Command[Literal["call_model", "agent_builder"]]:
+            """Execute tool calls"""
+            last_message = state["messages"][-1]
+            tool_calls = last_message.tool_calls if isinstance(last_message, AIMessage) else []
+
+            tool_messages = []
+            new_tool_ids = []
+            tool_result = ""
+            ask_user = False
+            ai_msg = ""
+            effective_previous_add_context = state.get("add_context", {})
+            effective_existing_context = state.get("context", {})
+            # logging.info(f"Initial new_tool_ids_for_context: {new_tool_ids_for_context}")
+
+            for tool_call in tool_calls:
+                tool_name = tool_call["name"]
+                tool_args = tool_call["args"]
+                try:
+                    if tool_name == "enter_agent_builder_mode":
+                        tool_message = ToolMessage(
+                            content=json.dumps("Entered Agent Builder Mode."),
+                            name=tool_call["name"],
+                            tool_call_id=tool_call["id"],
+                        )
+                        return Command(
+                            goto="agent_builder",
+                            update={
+                                "agent_builder_mode": "planning",
+                                "messages": [tool_message],
+                            },  # Entered Agent Builder mode
+                        )
+                    elif tool_name == "execute_ipython_cell":
+                        code = tool_call["args"]["snippet"]
+                        output, new_context, new_add_context = await handle_execute_ipython_cell(
+                            code,
+                            self.tools_context,  # Uses the dynamically updated context
+                            self.eval_fn,
+                            effective_previous_add_context,
+                            effective_existing_context,
+                        )
+                        effective_existing_context = new_context
+                        effective_previous_add_context = new_add_context
+                        tool_result = output
+                    elif tool_name == "load_functions":
+                        # The tool now does all the work of validation and formatting.
+                        tool_result, new_context_for_sandbox, valid_tools, unconnected_links = await meta_tools[
+                            "load_functions"
+                        ].ainvoke(tool_args)
+                        # We still need to update the sandbox context for `execute_ipython_cell`
+                        new_tool_ids.extend(valid_tools)
+                        if new_tool_ids:
+                            self.tools_context.update(new_context_for_sandbox)
+                        if unconnected_links:
+                            ask_user = True
+                            ai_msg = f"Please login to the following app(s) using the following links and let me know in order to proceed:\n {unconnected_links} "
+
+                    elif tool_name == "search_functions":
+                        tool_result = await meta_tools["search_functions"].ainvoke(tool_args)
+                    else:
+                        raise Exception(
+                            f"Unexpected tool call: {tool_call['name']}. "
+                            "tool calls must be one of 'enter_agent_builder_mode', 'execute_ipython_cell', 'load_functions', or 'search_functions'. For using functions, call them in code using 'execute_ipython_cell'."
+                        )
+                except Exception as e:
+                    tool_result = str(e)
+
+                tool_message = ToolMessage(
+                    content=json.dumps(tool_result),
+                    name=tool_call["name"],
+                    tool_call_id=tool_call["id"],
+                )
+                tool_messages.append(tool_message)
+
+            if ask_user:
+                tool_messages.append(AIMessage(content=ai_msg))
+                return Command(
+                    update={
+                        "messages": tool_messages,
+                        "selected_tool_ids": new_tool_ids,
+                        "context": effective_existing_context,
+                        "add_context": effective_previous_add_context,
+                    }
+                )
+
+            return Command(
+                goto="call_model",
+                update={
+                    "messages": tool_messages,
+                    "selected_tool_ids": new_tool_ids,
+                    "context": effective_existing_context,
+                    "add_context": effective_previous_add_context,
+                },
+            )
+
+        async def agent_builder(state: CodeActState, writer: StreamWriter) -> Command[Literal["call_model"]]:
+            agent_builder_mode = state.get("agent_builder_mode")
+            if agent_builder_mode == "planning":
+                plan_id = str(uuid.uuid4())
+                writer({"type": "custom", id: plan_id, "name": "planning", "data": {"update": bool(self.agent)}})
+                planning_instructions = self.instructions + AGENT_BUILDER_PLANNING_PROMPT + self.preloaded_defs
+                messages = [{"role": "system", "content": planning_instructions}] + strip_thinking(state["messages"])
+
+                model_with_structured_output = self.agent_builder_model_instance.with_structured_output(
+                    AgentBuilderPlan
+                )
+                response = await model_with_structured_output.ainvoke(messages)
+                plan = cast(AgentBuilderPlan, response)
+
+                writer({"type": "custom", id: plan_id, "name": "planning", "data": {"plan": plan.steps}})
+                ai_msg = AIMessage(
+                    content=json.dumps(plan.model_dump()),
+                    additional_kwargs={
+                        "type": "planning",
+                        "plan": plan.steps,
+                        "update": bool(self.agent),
+                    },
+                )
+
+                if self.eval_mode:
+                    mock_user_message = HumanMessage(content="yes, this is great")
+                    return Command(
+                        goto="agent_builder",
+                        update={
+                            "messages": [ai_msg, mock_user_message],
+                            "agent_builder_mode": "generating",
+                            "plan": plan.steps,
+                        },
+                    )
 
                 return Command(
-                    goto="call_model",
                     update={
-                        "messages": [
-                            ToolMessage(
-                                content=output,
-                                name=tool_call["name"],
-                                tool_call_id=tool_call["id"],
-                            )
-                        ],
-                        "context": new_context,
-                        "add_context": new_add_context,
+                        "messages": [ai_msg],
+                        "agent_builder_mode": "confirming",
+                        "plan": plan.steps,
+                    }
+                )
+
+            elif agent_builder_mode == "confirming":
+                # Deterministic routing based on three exact button inputs from UI
+                user_text = ""
+                for m in reversed(state["messages"]):
+                    try:
+                        if getattr(m, "type", "") in {"human", "user"}:
+                            user_text = (get_message_text(m) or "").strip()
+                            if user_text:
+                                break
+                    except Exception:
+                        continue
+
+                t = user_text.lower()
+                if t == "yes, this is great":
+                    self.meta_id = str(uuid.uuid4())
+                    name, description = None, None
+                    if self.agent:
+                        # Update flow: use existing name/description and do not re-generate
+                        name = getattr(self.agent, "name", None)
+                        description = getattr(self.agent, "description", None)
+                        writer(
+                            {
+                                "type": "custom",
+                                id: self.meta_id,
+                                "name": "generating",
+                                "data": {
+                                    "update": True,
+                                    "name": name,
+                                    "description": description,
+                                },
+                            }
+                        )
+                    else:
+                        writer({"type": "custom", id: self.meta_id, "name": "generating", "data": {"update": False}})
+
+                        meta_instructions = self.instructions + AGENT_BUILDER_META_PROMPT
+                        messages = [{"role": "system", "content": meta_instructions}] + state["messages"]
+
+                        model_with_structured_output = self.agent_builder_model_instance.with_structured_output(
+                            AgentBuilderMeta
+                        )
+                        meta_response = await model_with_structured_output.ainvoke(messages)
+                        meta = cast(AgentBuilderMeta, meta_response)
+                        name, description = meta.name, meta.description
+
+                        # Emit intermediary UI update with created name/description
+                        writer(
+                            {
+                                "type": "custom",
+                                id: self.meta_id,
+                                "name": "generating",
+                                "data": {"update": False, "name": name, "description": description},
+                            }
+                        )
+
+                    return Command(
+                        goto="agent_builder",
+                        update={
+                            "agent_builder_mode": "generating",
+                            "agent_name": name,
+                            "agent_description": description,
+                        },
+                    )
+                if t == "i would like to modify the plan":
+                    prompt_ai = AIMessage(
+                        content="What would you like to change about the plan? Let me know and I'll update the plan accordingly.",
+                        additional_kwargs={"stream": "true"},
+                    )
+                    return Command(update={"agent_builder_mode": "planning", "messages": [prompt_ai]})
+                if t == "let's do something else":
+                    return Command(goto="call_model", update={"agent_builder_mode": "inactive"})
+
+                # Fallback safe default
+                return Command(goto="call_model", update={"agent_builder_mode": "inactive"})
+
+            elif agent_builder_mode == "generating":
+                generating_instructions = self.instructions + AGENT_BUILDER_GENERATING_PROMPT + self.preloaded_defs
+                messages = [{"role": "system", "content": generating_instructions}] + state["messages"]
+
+                model_with_structured_output = self.agent_builder_model_instance.with_structured_output(
+                    AgentBuilderCode
+                )
+                response = await model_with_structured_output.ainvoke(messages)
+                func_code = cast(AgentBuilderCode, response).code
+
+                # Extract function name (handle both regular and async functions)
+                match = re.search(r"^\s*(?:async\s+)?def\s+([a-zA-Z_][a-zA-Z0-9_]*)\s*\(", func_code, re.MULTILINE)
+                if match:
+                    function_name = match.group(1)
+                else:
+                    function_name = "generated_agent"
+
+                # Use generated metadata if available
+                final_name = state.get("agent_name") or function_name
+                final_description = state.get("agent_description") or f"Generated agent: {function_name}"
+                add_context = state.get("add_context", {})
+                if "functions" not in add_context:
+                    add_context["functions"] = []
+                add_context["functions"].append(func_code)
+
+                # Save or update an Agent using the helper registry
+                try:
+                    if not self.agent_builder_registry:
+                        raise ValueError("AgentBuilder registry is not configured")
+
+                    plan_params = extract_plan_parameters(state["plan"])
+                    
+                    # Build instructions payload embedding the plan and function code
+                    instructions_payload = {
+                        "plan": state["plan"],
+                        "script": func_code,
+                        "params": plan_params,
+                    }
+
+                    # Convert tool ids list to dict
+                    tool_dict = convert_tool_ids_to_dict(state["selected_tool_ids"])
+
+                    res = self.agent_builder_registry.upsert_agent(
+                        name=final_name,
+                        description=final_description,
+                        instructions=instructions_payload,
+                        tools=tool_dict,
+                    )
+                except Exception:
+                    # In case of error, add the code to the exit message content
+
+                    mock_exit_tool_call = {"name": "exit_agent_builder_mode", "args": {}, "id": "exit_builder_1"}
+
+                    # Create a minimal assistant message to maintain flow
+                    mock_assistant_message = AIMessage(
+                        content=json.dumps(response.model_dump()),
+                        tool_calls=[mock_exit_tool_call],
+                        additional_kwargs={
+                            "type": "generating",
+                            "id": "ignore",
+                            "update": bool(self.agent),
+                            "name": final_name.replace(" ", "_"),
+                            "description": final_description,
+                        },
+                    )
+                    mock_exit_tool_response = ToolMessage(
+                        content=json.dumps(
+                            f"An error occurred. Displaying the function code:\n\n{func_code}\nFinal Name: {final_name}\nDescription: {final_description}"
+                        ),
+                        name="exit_agent_builder_mode",
+                        tool_call_id="exit_builder_1",
+                    )
+                    if self.eval_mode:
+                        human_msg = HumanMessage(content="Call the generated agent function (without redeclaring it) and check whether it works as expected")
+                        return Command(goto="call_model", update={"messages": [mock_assistant_message, mock_exit_tool_response, human_msg], "agent_builder_mode": "normal", "add_context": add_context})
+                    else:
+                        return Command(update={"messages": [mock_assistant_message, mock_exit_tool_response], "agent_builder_mode": "normal", "add_context": add_context})
+
+                writer(
+                    {
+                        "type": "custom",
+                        id: self.meta_id,
+                        "name": "generating",
+                        "data": {
+                            "id": str(res.id),
+                            "update": bool(self.agent),
+                            "name": final_name,
+                            "description": final_description,
+                            "add_context": add_context,
+                        },
+                    }
+                )
+                mock_exit_tool_call = {"name": "exit_agent_builder_mode", "args": {}, "id": "exit_builder_1"}
+                mock_assistant_message = AIMessage(
+                    content=json.dumps(response.model_dump()),
+                    tool_calls=[mock_exit_tool_call],
+                    additional_kwargs={
+                        "type": "generating",
+                        "id": str(res.id),
+                        "update": bool(self.agent),
+                        "name": final_name.replace(" ", "_"),
+                        "description": final_description,
                     },
                 )
 
+                mock_exit_tool_response = ToolMessage(
+                    content=json.dumps(
+                        "Exited Agent Builder Mode. Enter this mode again if you need to modify the saved agent."
+                    ),
+                    name="exit_agent_builder_mode",
+                    tool_call_id="exit_builder_1",
+                )
+
+                return Command(
+                    update={
+                        "messages": [mock_assistant_message, mock_exit_tool_response],
+                        "agent_builder_mode": "normal",
+                        "add_context": add_context,
+                    }
+                )
+
+        async def route_entry(state: CodeActState) -> Command[Literal["call_model", "agent_builder", "execute_tools"]]:
+            """Route to either normal mode or agent builder creation"""
+            pre_tools = await self.registry.export_tools(format=ToolFormat.NATIVE)
+
+            # Create the initial system prompt and tools_context in one go
+            self.final_instructions, self.tools_context = create_default_prompt(
+                pre_tools,
+                self.additional_tools,
+                self.instructions,
+                await get_connected_apps_string(self.registry),
+                self.agent,
+                is_initial_prompt=True,
+            )
+            self.preloaded_defs, _ = build_tool_definitions(pre_tools)
+            self.preloaded_defs = "\n".join(self.preloaded_defs)
+            await self.registry.load_tools(state["selected_tool_ids"])
+            exported_tools = await self.registry.export_tools(
+                state["selected_tool_ids"], ToolFormat.NATIVE
+            )  # Get definition for only the new tools
+            _, loaded_tools_context = build_tool_definitions(exported_tools)
+            self.tools_context.update(loaded_tools_context)
+
+            if (
+                len(state["messages"]) == 1 and self.agent
+            ):  # Inject the agent's script function into add_context for execution
+                script = self.agent.instructions.get("script")
+                add_context = {"functions": [script]}
+                return Command(goto="call_model", update={"add_context": add_context})
+
+            if state.get("agent_builder_mode") in ["planning", "confirming", "generating"]:
+                return Command(goto="agent_builder")
+            return Command(goto="call_model")
+
         agent = StateGraph(state_schema=CodeActState)
         agent.add_node(call_model, retry_policy=RetryPolicy(max_attempts=3, retry_on=filter_retry_on))
-        agent.add_node(sandbox)
-        agent.add_edge(START, "call_model")
+        agent.add_node(agent_builder)
+        agent.add_node(execute_tools)
+        agent.add_node(route_entry)
+        agent.add_edge(START, "route_entry")
         return agent.compile(checkpointer=self.memory)

universal_mcp/agents/codeact0/langgraph_agent.py

@@ -1,6 +1,6 @@
 from universal_mcp.agentr.registry import AgentrRegistry
 
-from universal_mcp.agents.codeact0.playbook_agent import CodeActPlaybookAgent
+from universal_mcp.agents.codeact0 import CodeActPlaybookAgent
 
 
 async def agent():