feat: add SDK command action helpers

.github/ISSUE_TEMPLATE/bug_report.yml

@@ -0,0 +1,65 @@
+name: Bug Report
+description: Report a reproducible problem or unexpected error in OpenOrbit
+title: "[Bug]: "
+labels: ["bug"]
+body:
+  - type: markdown
+    attributes:
+      value: |
+        Thanks for taking the time to report an issue! Please provide as much detail as possible to help us investigate and resolve it quickly.
+  - type: textarea
+    id: description
+    attributes:
+      label: Problem Description
+      description: A clear and concise description of the bug.
+      placeholder: Describe what happened...
+    validations:
+      required: true
+  - type: textarea
+    id: steps
+    attributes:
+      label: Steps to Reproduce
+      description: Exact steps to reproduce the behavior.
+      placeholder: |
+        1. Start the application with '...'
+        2. Navigate to '...'
+        3. Click on '...'
+        4. Observe error
+    validations:
+      required: true
+  - type: textarea
+    id: expected
+    attributes:
+      label: Expected Behavior
+      description: A clear and concise description of what you expected to happen.
+      placeholder: What did you expect to happen?
+    validations:
+      required: true
+  - type: textarea
+    id: actual
+    attributes:
+      label: Actual Behavior
+      description: What actually occurred, including any error messages or unexpected states.
+      placeholder: What happened instead?
+    validations:
+      required: true
+  - type: textarea
+    id: environment
+    attributes:
+      label: Environment & Versions
+      description: Relevant environment information (OS, Node version, Python version, browser, OpenOrbit version).
+      placeholder: |
+        - OS: [e.g. Linux Ubuntu 22.04 / macOS 14 / Windows 11]
+        - Python version: [e.g. 3.13.0]
+        - Node.js version: [e.g. 24.0.0]
+        - OpenOrbit version: [e.g. v0.8.0 or commit hash]
+    validations:
+      required: true
+  - type: textarea
+    id: additional-context
+    attributes:
+      label: Additional Context
+      description: Add any other context, logs, terminal outputs, or screenshots about the problem here.
+      placeholder: Optional additional logs or screenshots...
+    validations:
+      required: false

.github/ISSUE_TEMPLATE/feature_request.yml

@@ -0,0 +1,41 @@
+name: Feature Request
+description: Propose a new idea, enhancement, or capability for OpenOrbit
+title: "[Feature]: "
+labels: ["enhancement"]
+body:
+  - type: markdown
+    attributes:
+      value: |
+        Thank you for suggesting a feature! Please provide details about what problem this feature solves and how it should work.
+  - type: textarea
+    id: problem
+    attributes:
+      label: Problem Statement / Motivation
+      description: Is your feature request related to a problem or limitation? Please describe.
+      placeholder: A clear and concise description of what the problem or use case is...
+    validations:
+      required: true
+  - type: textarea
+    id: proposed-solution
+    attributes:
+      label: Proposed Solution & Expected Behavior
+      description: A clear and concise description of what you want to happen and how it should look or behave.
+      placeholder: Describe your proposed design, UI behavior, or API capability...
+    validations:
+      required: true
+  - type: textarea
+    id: alternatives
+    attributes:
+      label: Alternatives Considered
+      description: Any alternative solutions or workarounds you have considered.
+      placeholder: Optional description of other approaches evaluated...
+    validations:
+      required: false
+  - type: textarea
+    id: additional-context
+    attributes:
+      label: Additional Context
+      description: Add any other context, mockup screenshots, or implementation suggestions here.
+      placeholder: Optional mockups, architectural notes, or links...
+    validations:
+      required: false

.gitignore

@@ -16,3 +16,4 @@ test-results/
 openorbit-state.zip
 conversations/
 .orbit/
+AGENTS.md

README.md

@@ -278,6 +278,7 @@ OpenOrbit exposes a local, versioned API:
 - Swagger UI: `http://localhost:3000/api/docs`
 - OpenAPI document: `http://localhost:3000/api/openapi.json`
 - API base: `http://localhost:3000/api/v1`
+- MCP (Streamable HTTP): `http://localhost:3000/mcp/`
 
 Read the [API reference](docs/API.md) for endpoint details. To add reusable automation, create a Python runner with explicit lifecycle phases:
 

backend/app/assistant_graph.py

@@ -0,0 +1,159 @@
+"""LangGraph orchestration for the local Orbit assistant."""
+
+from __future__ import annotations
+
+import asyncio
+import json
+from typing import Any, Callable, TypedDict
+
+from coding_agents import execution_environment_context
+from langgraph.graph import END, START, StateGraph
+from mcp import ClientSession
+from mcp.client.streamable_http import streamablehttp_client
+
+from .assistant_tools import DEFAULT_MCP_URL, AssistantToolExecutor
+from .assistant_ui import AssistantUiToolExecutor
+from .providers import AzureOpenAIProvider, BedrockProvider, ModelSettings
+
+
+class AssistantState(TypedDict, total=False):
+    prompt: str
+    response: str
+
+
+class LocalMcpTools:
+    """Expose the mounted OpenOrbit MCP server as native model function tools."""
+
+    def __init__(self, url: str = DEFAULT_MCP_URL):
+        self.url = url
+
+    async def _list_tools(self) -> list[dict[str, Any]]:
+        async with streamablehttp_client(self.url) as (read, write, _):
+            async with ClientSession(read, write) as session:
+                await session.initialize()
+                result = await session.list_tools()
+                return [
+                    {
+                        "name": tool.name,
+                        "description": tool.description or "OpenOrbit MCP tool.",
+                        "parameters": tool.inputSchema,
+                    }
+                    for tool in result.tools
+                ]
+
+    async def _call_tool(self, name: str, arguments: dict[str, Any]) -> str:
+        async with streamablehttp_client(self.url) as (read, write, _):
+            async with ClientSession(read, write) as session:
+                await session.initialize()
+                result = await session.call_tool(name, arguments)
+                if result.isError:
+                    return json.dumps({"error": "MCP tool failed", "content": result.content}, default=str)
+                return json.dumps({"content": result.content}, default=str, ensure_ascii=False)
+
+    @staticmethod
+    def _run(coro: Any) -> Any:
+        return asyncio.run(coro)
+
+    def definitions(self) -> list[dict[str, Any]]:
+        try:
+            return self._run(self._list_tools())
+        except Exception:
+            # The assistant remains usable while the local server is starting.
+            return []
+
+    def execute(self, name: str, arguments: dict[str, Any]) -> str:
+        try:
+            return self._run(self._call_tool(name, arguments))
+        except Exception as error:
+            return json.dumps({"error": f"MCP unavailable: {error}"})
+
+
+class OrbitAssistantGraph:
+    """A small, explicit LangGraph for one bounded assistant turn.
+
+    The model node can use both the local bounded workspace tools and the
+    mounted MCP control-room tools. Provider adapters retain the native tool
+    call protocol, while LangGraph owns the turn state and final response.
+    """
+
+    def __init__(
+        self,
+        provider: AzureOpenAIProvider | BedrockProvider,
+        settings: ModelSettings,
+        local_tools: AssistantToolExecutor,
+        mcp_tools: LocalMcpTools | None = None,
+        ui_tools: AssistantUiToolExecutor | None = None,
+        on_activity: Callable[[str, str | None], None] | None = None,
+    ):
+        self.provider = provider
+        self.settings = settings
+        self.local_tools = local_tools
+        self.mcp_tools = mcp_tools or LocalMcpTools(local_tools.settings["mcp_server_url"])
+        self.ui_tools = ui_tools
+        self.on_activity = on_activity
+        graph = StateGraph(AssistantState)
+        graph.add_node("model", self._model)
+        graph.add_edge(START, "model")
+        graph.add_edge("model", END)
+        self.graph = graph.compile()
+
+    def _model(self, state: AssistantState) -> AssistantState:
+        self._activity("thinking")
+        local_definitions = self.local_tools.definitions()
+        mcp_definitions = self.mcp_tools.definitions()
+        ui_definitions = self.ui_tools.definitions() if self.ui_tools else []
+        definitions = [*mcp_definitions, *local_definitions, *ui_definitions]
+
+        def execute(name: str, arguments: dict[str, Any]) -> str:
+            self._activity("working", name)
+            if any(tool["name"] == name for tool in mcp_definitions):
+                result = self.mcp_tools.execute(name, arguments)
+            elif self.ui_tools and any(tool["name"] == name for tool in ui_definitions):
+                result = self.ui_tools.execute(name, arguments)
+            else:
+                result = self.local_tools.execute(name, arguments)
+            self._activity("thinking")
+            return result
+
+        response = (
+            self.provider.complete_with_tools(self.settings, state["prompt"], definitions, execute)
+            if definitions
+            else self.provider.complete(self.settings, state["prompt"])
+        )
+        return {"response": response}
+
+    def _activity(self, phase: str, tool: str | None = None) -> None:
+        if self.on_activity:
+            self.on_activity(phase, tool)
+
+    def invoke(self, prompt: str) -> str:
+        result = self.graph.invoke({"prompt": prompt})
+        return str(result["response"])
+
+
+def build_assistant_prompt(
+    content: str, history: list[tuple[str, str]], locale: str | None, ui_enabled: bool = False
+) -> str:
+    prompt = (
+        "You are Orbit, a concise assistant for the local OpenOrbit control room.\n"
+        "Use the OpenOrbit MCP tools as the source of truth for local control-room state. "
+        "Read state before suggesting or taking any state-changing action.\n"
+        + execution_environment_context()
+    )
+    if ui_enabled:
+        prompt += (
+            "Browser UI tools are available for this turn. For any question about the current Orbit "
+            "screen, visible component, selected item, open panel, dialog, form value, or any request "
+            "to change the Orbit UI, first call ui_get_context. Use ui_interact only after reading the "
+            "current UI context. Do not use file tools to infer browser UI state. File tools remain for "
+            "workspace files and coding tasks only.\n"
+        )
+    if locale:
+        prompt += f"Respond in BCP 47 locale '{locale}'.\n"
+    if history:
+        prompt += (
+            "Conversation so far:\n"
+            + "\n".join(f"{role.title()}: {message}" for role, message in history)
+            + "\n\n"
+        )
+    return prompt + f"User: {content}\nAssistant:"