wcgw 1.5.1__py3-none-any.whl → 1.5.3__py3-none-any.whl

wcgw/client/anthropic_client.py

@@ -182,6 +182,9 @@ def loop(
 - Send text input to the running program.
 - Send send_specials=["Enter"] to recheck status of a running program.
 - Only one of send_text, send_specials, send_ascii should be provided.
+- This returns within 5 seconds, for heavy programs keep checking status for upto 10 turns before asking user to continue checking again.
+    - Programs don't hang easily, so most likely explanation for no output is usually that the program is still running, and you need to check status again usign ["Enter"].
+
 """,
         ),
         ToolParam(
@@ -223,9 +226,10 @@ def loop(
             input_schema=GetScreenInfo.model_json_schema(),
             name="GetScreenInfo",
             description="""
-- Get display information of an OS running on docker using image "ghcr.io/anthropics/anthropic-quickstarts:computer-use-demo-latest"
-- If user hasn't provided docker image id, check using `docker ps` and provide the id.
 - Important: call this first in the conversation before ScreenShot, Mouse, and Keyboard tools.
+- Get display information of a linux os running on docker using image "ghcr.io/anthropics/anthropic-quickstarts:computer-use-demo-latest"
+- If user hasn't provided docker image id, check using `docker ps` and provide the id.
+- If the docker is not running, run using `docker run -d -p 6080:6080 ghcr.io/anthropics/anthropic-quickstarts:computer-use-demo-latest`
 - Connects shell to the docker environment.
 - Note: once this is called, the shell enters the docker environment. All bash commands will run over there.
 """,
@@ -234,26 +238,29 @@ def loop(
             input_schema=ScreenShot.model_json_schema(),
             name="ScreenShot",
             description="""
-- Capture screenshot of an OS running on docker using image "ghcr.io/anthropics/anthropic-quickstarts:computer-use-demo-latest"
-- If user hasn't provided docker image id, check using `docker ps` and provide the id.
-- Capture ScreenShot of the current screen for automation.
+- Capture screenshot of the linux os on docker.
+- All actions on UI using mouse and keyboard return within 0.5 seconds.
+    * So if you're doing something that takes longer for UI to update like heavy page loading, keep checking UI for update usign ScreenShot upto 10 turns. 
+    * Notice for smallest of the loading icons to check if your action worked.
+    * After 10 turns of no change, ask user for permission to keep checking.
+    * If you don't notice even slightest of the change, it's likely you clicked on the wrong place.
+
 """,
         ),
         ToolParam(
             input_schema=Mouse.model_json_schema(),
             name="Mouse",
             description="""
-- Interact with docker container running image "ghcr.io/anthropics/anthropic-quickstarts:computer-use-demo-latest"
-- If user hasn't provided docker image id, check using `docker ps` and provide the id.
-- Interact with the screen using mouse
+- Interact with the linux os on docker using mouse.
+- Uses xdotool
+- About left_click_drag: the current mouse position will be used as the starting point, click and drag to the given x, y coordinates. Useful in things like sliders, moving things around, etc.
 """,
         ),
         ToolParam(
             input_schema=Keyboard.model_json_schema(),
             name="Keyboard",
             description="""
-- Interact with docker container running image "ghcr.io/anthropics/anthropic-quickstarts:computer-use-demo-latest"
-- If user hasn't provided docker image id, check using `docker ps` and provide the id.
+- Interact with the linux os on docker using keyboard.
 - Emulate keyboard input to the screen
 - Uses xdootool to send keyboard input, keys like Return, BackSpace, Escape, Page_Up, etc. can be used.
 - Do not use it to interact with Bash tool.

wcgw/client/computer_use.py

@@ -26,6 +26,7 @@ from ..types_ import (
 OUTPUT_DIR = "/tmp/outputs"
 TYPING_DELAY_MS = 12
 TYPING_GROUP_SIZE = 50
+SLEEP_TIME_MAX_S = 3
 
 Action = Literal[
     "key",
@@ -186,12 +187,16 @@ class ComputerTool:
         docker_image_id: Optional[str] = None,
         text: str | None = None,
         coordinate: tuple[int, int] | None = None,
+        do_left_click_on_move: bool | None = None,
+        take_after_delay_seconds: int | None = None,
         **kwargs: Any,
     ) -> ToolResult:
         if action == "get_screen_info":
             assert docker_image_id is not None
             self.docker_image_id = docker_image_id
             self.get_screen_info()
+            if take_after_delay_seconds is not None:
+                time.sleep(min(take_after_delay_seconds, SLEEP_TIME_MAX_S))
             screenshot_res = self.screenshot()
             return ToolResult(
                 output=f"width: {self.width}, height: {self.height}, display_num: {self.display_num}",
@@ -217,7 +222,12 @@ class ComputerTool:
             )
 
             if action == "mouse_move":
-                return self.shell(f"{self.xdotool} mousemove  {x} {y}")
+                if not do_left_click_on_move:
+                    return self.shell(f"{self.xdotool} mousemove  {x} {y}")
+                else:
+                    return self.shell(
+                        f"{self.xdotool} mousemove  {x} {y} click 1",
+                    )
             elif action == "left_click_drag":
                 return self.shell(
                     f"{self.xdotool} mousedown 1 mousemove  {x} {y} mouseup 1",
@@ -390,6 +400,7 @@ def run_computer_tool(
     elif isinstance(action, ScreenShot):
         result = Computer(
             action="screenshot",
+            screenshot_delay=action.take_after_delay_seconds,
         )
     elif isinstance(action, Keyboard):
         result = Computer(
@@ -401,6 +412,7 @@ def run_computer_tool(
             result = Computer(
                 action="mouse_move",
                 coordinate=(action.action.x, action.action.y),
+                do_left_click_on_move=action.action.do_left_click_on_move,
             )
         elif isinstance(action.action, LeftClickDrag):
             result = Computer(

wcgw/client/mcp_server/Readme.md

@@ -1,5 +1,9 @@
 # Claude desktop support
 
+`wcgw` enables Claude desktop app on Mac to access shell and file system in order to automate tasks, run code, etc.
+
+It also has a computer use feature to connect to linux running on docker. Claude can fully control it including mouse and keyboard.
+
 ## Setup
 
 Update `claude_desktop_config.json` (~/Library/Application Support/Claude/claude_desktop_config.json)
@@ -32,21 +36,10 @@ Computer use is enabled by default. Claude will be able to connect to any docker
 First run a sample docker image with desktop and optionally VNC connection:
 
 ```sh
-docker run \
-    --entrypoint "" \
-    -p 6080:6080 \
-    -e WIDTH=1024 \
-    -e HEIGHT=768 \
-    -d \
-    ghcr.io/anthropics/anthropic-quickstarts:computer-use-demo-latest \
-    bash -c "\
-        ./start_all.sh && \
-        ./novnc_startup.sh && \
-        python http_server.py > /tmp/server_logs.txt 2>&1 & \
-        tail -f /dev/null"
+docker run -p 6080:6080 ghcr.io/anthropics/anthropic-quickstarts:computer-use-demo-latest
 ```
 
-Connect to `http://localhost:6080/vnc.html` for desktop view (VNC) of the system running in the docker. Then ask claude to control the docker os.
+Connect to `http://localhost:6080/vnc.html` for desktop view (VNC) of the system running in the docker. Then ask claude desktop app to control the docker os.
 
 ## Usage
 

wcgw/client/mcp_server/server.py

@@ -28,9 +28,9 @@ from ...types_ import (
     ScreenShot,
     GetScreenInfo,
 )
-from ..computer_use import Computer
+from ..computer_use import SLEEP_TIME_MAX_S
 
-tools.TIMEOUT = 3
+tools.TIMEOUT = SLEEP_TIME_MAX_S
 
 server = Server("wcgw")
 
@@ -102,6 +102,8 @@ async def handle_list_tools() -> list[types.Tool]:
 - Send text input to the running program.
 - Send send_specials=["Enter"] to recheck status of a running program.
 - Only one of send_text, send_specials, send_ascii should be provided.
+- This returns within 3 seconds, for heavy programs keep checking status for upto 10 turns before asking user to continue checking again.
+    - Programs don't hang easily, so most likely explanation for no output is usually that the program is still running, and you need to check status again usign ["Enter"].
 """,
         ),
         ToolParam(
@@ -151,9 +153,10 @@ async def handle_list_tools() -> list[types.Tool]:
             inputSchema=GetScreenInfo.model_json_schema(),
             name="GetScreenInfo",
             description="""
-- Get display information of an OS running on docker using image "ghcr.io/anthropics/anthropic-quickstarts:computer-use-demo-latest"
-- If user hasn't provided docker image id, check using `docker ps` and provide the id.
 - Important: call this first in the conversation before ScreenShot, Mouse, and Keyboard tools.
+- Get display information of a linux os running on docker using image "ghcr.io/anthropics/anthropic-quickstarts:computer-use-demo-latest"
+- If user hasn't provided docker image id, check using `docker ps` and provide the id.
+- If the docker is not running, run using `docker run -d -p 6080:6080 ghcr.io/anthropics/anthropic-quickstarts:computer-use-demo-latest`
 - Connects shell to the docker environment.
 - Note: once this is called, the shell enters the docker environment. All bash commands will run over there.
 """,
@@ -162,29 +165,32 @@ async def handle_list_tools() -> list[types.Tool]:
             inputSchema=ScreenShot.model_json_schema(),
             name="ScreenShot",
             description="""
-- Capture screenshot of an OS running on docker using image "ghcr.io/anthropics/anthropic-quickstarts:computer-use-demo-latest"
-- If user hasn't provided docker image id, check using `docker ps` and provide the id.
-- Capture ScreenShot of the current screen for automation.
+- Capture screenshot of the linux os on docker.
+- All actions on UI using mouse and keyboard return within 0.5 seconds.
+    * So if you're doing something that takes longer for UI to update like heavy page loading, keep checking UI for update usign ScreenShot upto 10 turns. 
+    * Notice for smallest of the loading icons to check if your action worked.
+    * After 10 turns of no change, ask user for permission to keep checking.
+    * If you don't notice even slightest of the change, it's likely you clicked on the wrong place.
 """,
         ),
         ToolParam(
             inputSchema=Mouse.model_json_schema(),
             name="Mouse",
             description="""
-- Interact with docker container running image "ghcr.io/anthropics/anthropic-quickstarts:computer-use-demo-latest"
-- If user hasn't provided docker image id, check using `docker ps` and provide the id.
-- Interact with the screen using mouse
+- Interact with the linux os on docker using mouse.
+- Uses xdotool
+- About left_click_drag: the current mouse position will be used as the starting point, click and drag to the given x, y coordinates. Useful in things like sliders, moving things around, etc.
 """,
         ),
         ToolParam(
             inputSchema=Keyboard.model_json_schema(),
             name="Keyboard",
             description="""
-- Interact with docker container running image "ghcr.io/anthropics/anthropic-quickstarts:computer-use-demo-latest"
-- If user hasn't provided docker image id, check using `docker ps` and provide the id.
+- Interact with the linux os on docker using keyboard.
 - Emulate keyboard input to the screen
 - Uses xdootool to send keyboard input, keys like Return, BackSpace, Escape, Page_Up, etc. can be used.
 - Do not use it to interact with Bash tool.
+- Make sure you've selected a text area or an editable element before sending text.
 """,
         ),
     ]
@@ -259,27 +265,18 @@ async def handle_call_tool(
 
 async def main() -> None:
     version = importlib.metadata.version("wcgw")
-    while True:
-        try:
-            # Run the server using stdin/stdout streams
-            async with mcp.server.stdio.stdio_server() as (read_stream, write_stream):
-                await server.run(
-                    read_stream,
-                    write_stream,
-                    InitializationOptions(
-                        server_name="wcgw",
-                        server_version=version,
-                        capabilities=server.get_capabilities(
-                            notification_options=NotificationOptions(),
-                            experimental_capabilities={},
-                        ),
-                    ),
-                    raise_exceptions=False,
-                )
-        except BaseException as e:
-            print(f"Server encountered an error: {e}", file=sys.stderr)
-            print("Stack trace:", file=sys.stderr)
-            traceback.print_exc(file=sys.stderr)
-            print("Restarting server in 5 seconds...", file=sys.stderr)
-            await asyncio.sleep(5)
-            continue
+    # Run the server using stdin/stdout streams
+    async with mcp.server.stdio.stdio_server() as (read_stream, write_stream):
+        await server.run(
+            read_stream,
+            write_stream,
+            InitializationOptions(
+                server_name="wcgw",
+                server_version=version,
+                capabilities=server.get_capabilities(
+                    notification_options=NotificationOptions(),
+                    experimental_capabilities={},
+                ),
+            ),
+            raise_exceptions=False,
+        )

wcgw/client/tools.py

@@ -170,6 +170,7 @@ def initial_info() -> str:
 System: {uname_sysname}
 Machine: {uname_machine}
 Current working directory: {CWD}
+wcgw version: {importlib.metadata.version("wcgw")}
 """
 
 

wcgw/types_.py

@@ -60,12 +60,13 @@ class GetScreenInfo(BaseModel):
 
 class ScreenShot(BaseModel):
     type: Literal["ScreenShot"]
-    docker_image_id: str
+    take_after_delay_seconds: int
 
 
 class MouseMove(BaseModel):
     x: int
     y: int
+    do_left_click_on_move: bool
     type: Literal["MouseMove"]
 
 

{wcgw-1.5.1.dist-info → wcgw-1.5.3.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: wcgw
-Version: 1.5.1
+Version: 1.5.3
 Summary: What could go wrong giving full shell access to chatgpt?
 Project-URL: Homepage, https://github.com/rusiaaman/wcgw
 Author-email: Aman Rusia <gapypi@arcfu.com>
@@ -29,7 +29,9 @@ Description-Content-Type: text/markdown
 
 # Shell and Coding agent on Chatgpt and Claude desktop apps
 
-A custom gpt on chatgpt web/desktop apps to interact with your local shell, edit files, run code, etc.
+- An MCP server on claude desktop for autonomous shell, coding and desktop control agent.
+- A custom gpt on chatgpt web/desktop apps to interact with your local shell, edit files, run code, etc.
+
 
 [![Tests](https://github.com/rusiaaman/wcgw/actions/workflows/python-tests.yml/badge.svg?branch=main)](https://github.com/rusiaaman/wcgw/actions/workflows/python-tests.yml)
 [![Build](https://github.com/rusiaaman/wcgw/actions/workflows/python-publish.yml/badge.svg)](https://github.com/rusiaaman/wcgw/actions/workflows/python-publish.yml)
@@ -40,6 +42,7 @@ A custom gpt on chatgpt web/desktop apps to interact with your local shell, edit
 ### 🚀 Highlights
 
 - ⚡ **Full Shell Access**: No restrictions, complete control.
+- ⚡ **Desktop control on Claude**: Screen capture, mouse control, keyboard control on claude desktop (on mac with docker linux)
 - ⚡ **Create, Execute, Iterate**: Ask the gpt to keep running compiler checks till all errors are fixed, or ask it to keep checking for the status of a long running command till it's done.
 - ⚡ **Interactive Command Handling**: Supports interactive commands using arrow keys, interrupt, and ansi escape sequences.
 - ⚡ **REPL support**: [beta] Supports python/node and other REPL execution.

{wcgw-1.5.1.dist-info → wcgw-1.5.3.dist-info}/RECORD

@@ -1,22 +1,22 @@
 wcgw/__init__.py,sha256=9K2QW7QuSLhMTVbKbBYd9UUp-ZyrfBrxcjuD_xk458k,118
-wcgw/types_.py,sha256=STKGeVdQNK_k8mcmAXpKZJY9YUpU7-mbJcFpXfzBuys,1732
+wcgw/types_.py,sha256=rDz4olJS2zvYC13jzeOppA2tci-tVDyWAqeA5BesAaU,1773
 wcgw/client/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 wcgw/client/__main__.py,sha256=wcCrL4PjG51r5wVKqJhcoJPTLfHW0wNbD31DrUN0MWI,28
-wcgw/client/anthropic_client.py,sha256=owR-nIxQVGgw_ned8JOQ-QmmCBQvZSqgD08kYpU_Rbg,17730
+wcgw/client/anthropic_client.py,sha256=c6nNNojsWLFWjMJcEsEp7g31Ps9SIgCqMk4dhrOn9V4,18314
 wcgw/client/cli.py,sha256=Oja42CHkVO8puqOXflko9NeephYCMa85aBmQTEjBZtI,932
 wcgw/client/common.py,sha256=grH-yV_4tnTQZ29xExn4YicGLxEq98z-HkEZwH0ReSg,1410
-wcgw/client/computer_use.py,sha256=hadmsHpwVRqTZh4Q7Ssu3xdnNLnBW4y-pd5P6D-qqKE,14276
+wcgw/client/computer_use.py,sha256=eGiINKfgY8WWT-NDUa6vUKd1MTWE7dTjSlvjZHPCWzc,14870
 wcgw/client/diff-instructions.txt,sha256=s5AJKG23JsjwRYhFZFQVvwDpF67vElawrmdXwvukR1A,1683
 wcgw/client/openai_client.py,sha256=L61ajFVQW2QPS3C0n1YsjgF4vQKfMIZHmp6iFBHutX8,17748
 wcgw/client/openai_utils.py,sha256=YNwCsA-Wqq7jWrxP0rfQmBTb1dI0s7dWXzQqyTzOZT4,2629
 wcgw/client/sys_utils.py,sha256=GajPntKhaTUMn6EOmopENWZNR2G_BJyuVbuot0x6veI,1376
-wcgw/client/tools.py,sha256=Ce_1eLXl6W1U2EcNk2JPiCAKmEnTHt3Jd78ZHlW-ET4,32629
-wcgw/client/mcp_server/Readme.md,sha256=mztipVTwqXLLenbVihnodq7gUF2Q0_YOKOMW3MiV3UM,2020
+wcgw/client/tools.py,sha256=d7Fni7JU3aOh2vXBAw5k5rsxkdQVcxoxc5vipvEsA2g,32680
+wcgw/client/mcp_server/Readme.md,sha256=1hNZtqltsORug7OzUjjoK5O8q5s9-Y3S0_rlzT-Wfg4,2033
 wcgw/client/mcp_server/__init__.py,sha256=cQ7PUrEmXUpio8x0SEoGWP5hCRPd7z2bAkNCbYbtTys,236
-wcgw/client/mcp_server/server.py,sha256=niS5elM7vdu181DjDSNXRbipFjt9Ke-H9HQMWMwNnXg,10211
+wcgw/client/mcp_server/server.py,sha256=i1nn16LSsbySm3LCt_T_D2G33nPtGNDJakp5FzKV6vQ,10416
 wcgw/relay/serve.py,sha256=RUcUeyL4Xt0EEo12Ul6VQjb4tRle4uIdsa85v7XXxEw,8771
 wcgw/relay/static/privacy.txt,sha256=s9qBdbx2SexCpC_z33sg16TptmAwDEehMCLz4L50JLc,529
-wcgw-1.5.1.dist-info/METADATA,sha256=ls5ebZhUUraGJqdSIHCggftaq16jnGJS8VxHMj1Y0GY,6285
-wcgw-1.5.1.dist-info/WHEEL,sha256=C2FUgwZgiLbznR-k0b_5k3Ai_1aASOXDss3lzCUsUug,87
-wcgw-1.5.1.dist-info/entry_points.txt,sha256=eKo1omwbAggWlQ0l7GKoR7uV1-j16nk9tK0BhC2Oz_E,120
-wcgw-1.5.1.dist-info/RECORD,,
+wcgw-1.5.3.dist-info/METADATA,sha256=9raP3OP6KWq5bzIEuwJ41_04g5fS5m1HVSNm-YI8Dkw,6508
+wcgw-1.5.3.dist-info/WHEEL,sha256=C2FUgwZgiLbznR-k0b_5k3Ai_1aASOXDss3lzCUsUug,87
+wcgw-1.5.3.dist-info/entry_points.txt,sha256=eKo1omwbAggWlQ0l7GKoR7uV1-j16nk9tK0BhC2Oz_E,120
+wcgw-1.5.3.dist-info/RECORD,,