The Windows CLI MCP Server enables secure command-line interactions on Windows systems and remote systems via SSH, with robust security controls and multi-shell support.
Execute commands in PowerShell, Command Prompt (CMD), and Git Bash within configured security constraints
Manage SSH connections to execute commands on remote systems (create, read, update, delete connections)
Access command history with timestamps and outputs
Get the current working directory of the server process
Enforce security with features like command blocking, argument blocking, working directory restriction, and injection protection
Configure SSH with detailed connection settings (host, port, username, password, private key)
Log commands for auditing and compliance purposes
Disconnect from active SSH sessions
Provides access to Git Bash shell for executing Git commands and scripts on Windows, allowing repository management and version control operations.
Supports SSH connections to Raspberry Pi devices for remote command execution, configured through the SSH connection profiles.
Enables controlled execution of shell commands across PowerShell, CMD, and Git Bash with security restrictions, command blocking, and history tracking.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@Windows CLI MCP Server list files in the current directory using PowerShell"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
PROJECT DEPRECATED - No longer maintained. Use https://github.com/wonderwhy-er/DesktopCommanderMCP instead for similar functionality.
MCP server for secure command-line interactions on Windows systems, enabling controlled access to PowerShell, CMD, Git Bash shells, and remote systems via SSH. It allows MCP clients (like Claude Desktop) to perform operations on your system, similar to Open Interpreter.
This MCP server provides direct access to your system's command line interface and remote systems via SSH. When enabled, it grants access to your files, environment variables, command execution capabilities, and remote server management.
Review and restrict allowed paths and SSH connections
Enable directory restrictions
Configure command blocks
Consider security implications
See Configuration for more details.
Multi-Shell Support: Execute commands in PowerShell, Command Prompt (CMD), and Git Bash
SSH Support: Execute commands on remote systems via SSH
Resource Exposure: View SSH connections, current directory, and configuration as MCP resources
Security Controls:
Command and SSH command blocking (full paths, case variations)
Working directory validation
Maximum command length limits
Command logging and history tracking
Smart argument validation
Configurable:
Custom security rules
Shell-specific settings
SSH connection profiles
Path restrictions
Blocked command lists
See the API section for more details on the tools and resources the server provides to MCP clients.
Note: The server will only allow operations within configured directories, with allowed commands, and on configured SSH connections.
Related MCP server: Super Shell MCP Server
Add this to your claude_desktop_config.json:
{
"mcpServers": {
"windows-cli": {
"command": "npx",
"args": ["-y", "@simonb97/server-win-cli"]
}
}
}For use with a specific config file, add the --config flag:
{
"mcpServers": {
"windows-cli": {
"command": "npx",
"args": [
"-y",
"@simonb97/server-win-cli",
"--config",
"path/to/your/config.json"
]
}
}
}After configuring, you can:
Execute commands directly using the available tools
View configured SSH connections and server configuration in the Resources section
Manage SSH connections through the provided tools
The server uses a JSON configuration file to customize its behavior. You can specify settings for security controls, shell configurations, and SSH connections.
To create a default config file, either:
a) copy config.json.example to config.json, or
b) run:
npx @simonb97/server-win-cli --init-config ./config.jsonThen set the --config flag to point to your config file as described in the Usage with Claude Desktop section.
The server looks for configuration in the following locations (in order):
Path specified by --config flag
./config.json in current directory
~/.win-cli-mcp/config.json in user's home directory
If no configuration file is found, the server will use a default (restricted) configuration:
Note: The default configuration is designed to be restrictive and secure. Find more details on each setting in the Configuration Settings section.
{
"security": {
"maxCommandLength": 2000,
"blockedCommands": [
"rm",
"del",
"rmdir",
"format",
"shutdown",
"restart",
"reg",
"regedit",
"net",
"netsh",
"takeown",
"icacls"
],
"blockedArguments": [
"--exec",
"-e",
"/c",
"-enc",
"-encodedcommand",
"-command",
"--interactive",
"-i",
"--login",
"--system"
],
"allowedPaths": ["User's home directory", "Current working directory"],
"restrictWorkingDirectory": true,
"logCommands": true,
"maxHistorySize": 1000,
"commandTimeout": 30,
"enableInjectionProtection": true
},
"shells": {
"powershell": {
"enabled": true,
"command": "powershell.exe",
"args": ["-NoProfile", "-NonInteractive", "-Command"],
"blockedOperators": ["&", "|", ";", "`"]
},
"cmd": {
"enabled": true,
"command": "cmd.exe",
"args": ["/c"],
"blockedOperators": ["&", "|", ";", "`"]
},
"gitbash": {
"enabled": true,
"command": "C:\\Program Files\\Git\\bin\\bash.exe",
"args": ["-c"],
"blockedOperators": ["&", "|", ";", "`"]
}
},
"ssh": {
"enabled": false,
"defaultTimeout": 30,
"maxConcurrentSessions": 5,
"keepaliveInterval": 10000,
"keepaliveCountMax": 3,
"readyTimeout": 20000,
"connections": {}
}
}The configuration file is divided into three main sections: security, shells, and ssh.
{
"security": {
// Maximum allowed length for any command
"maxCommandLength": 1000,
// Commands to block - blocks both direct use and full paths
// Example: "rm" blocks both "rm" and "C:\\Windows\\System32\\rm.exe"
// Case-insensitive: "del" blocks "DEL.EXE", "del.cmd", etc.
"blockedCommands": [
"rm", // Delete files
"del", // Delete files
"rmdir", // Delete directories
"format", // Format disks
"shutdown", // Shutdown system
"restart", // Restart system
"reg", // Registry editor
"regedit", // Registry editor
"net", // Network commands
"netsh", // Network commands
"takeown", // Take ownership of files
"icacls" // Change file permissions
],
// Arguments that will be blocked when used with any command
// Note: Checks each argument independently - "cd warm_dir" won't be blocked just because "rm" is in blockedCommands
"blockedArguments": [
"--exec", // Execution flags
"-e", // Short execution flags
"/c", // Command execution in some shells
"-enc", // PowerShell encoded commands
"-encodedcommand", // PowerShell encoded commands
"-command", // Direct PowerShell command execution
"--interactive", // Interactive mode which might bypass restrictions
"-i", // Short form of interactive
"--login", // Login shells might have different permissions
"--system" // System level operations
],
// List of directories where commands can be executed
"allowedPaths": ["C:\\Users\\YourUsername", "C:\\Projects"],
// If true, commands can only run in allowedPaths
"restrictWorkingDirectory": true,
// If true, saves command history
"logCommands": true,
// Maximum number of commands to keep in history
"maxHistorySize": 1000,
// Timeout for command execution in seconds (default: 30)
"commandTimeout": 30,
// Enable or disable protection against command injection (covers ;, &, |, \`)
"enableInjectionProtection": true
}
}{
"shells": {
"powershell": {
// Enable/disable this shell
"enabled": true,
// Path to shell executable
"command": "powershell.exe",
// Default arguments for the shell
"args": ["-NoProfile", "-NonInteractive", "-Command"],
// Optional: Specify which command operators to block
"blockedOperators": ["&", "|", ";", "`"] // Block all command chaining
},
"cmd": {
"enabled": true,
"command": "cmd.exe",
"args": ["/c"],
"blockedOperators": ["&", "|", ";", "`"] // Block all command chaining
},
"gitbash": {
"enabled": true,
"command": "C:\\Program Files\\Git\\bin\\bash.exe",
"args": ["-c"],
"blockedOperators": ["&", "|", ";", "`"] // Block all command chaining
}
}
}{
"ssh": {
// Enable/disable SSH functionality
"enabled": false,
// Default timeout for SSH commands in seconds
"defaultTimeout": 30,
// Maximum number of concurrent SSH sessions
"maxConcurrentSessions": 5,
// Interval for sending keepalive packets (in milliseconds)
"keepaliveInterval": 10000,
// Maximum number of failed keepalive attempts before disconnecting
"keepaliveCountMax": 3,
// Timeout for establishing SSH connections (in milliseconds)
"readyTimeout": 20000,
// SSH connection profiles
"connections": {
// NOTE: these examples are not set in the default config!
// Example: Local Raspberry Pi
"raspberry-pi": {
"host": "raspberrypi.local", // Hostname or IP address
"port": 22, // SSH port
"username": "pi", // SSH username
"password": "raspberry", // Password authentication (if not using key)
"keepaliveInterval": 10000, // Override global keepaliveInterval
"keepaliveCountMax": 3, // Override global keepaliveCountMax
"readyTimeout": 20000 // Override global readyTimeout
},
// Example: Remote server with key authentication
"dev-server": {
"host": "dev.example.com",
"port": 22,
"username": "admin",
"privateKeyPath": "C:\\Users\\YourUsername\\.ssh\\id_rsa", // Path to private key
"keepaliveInterval": 10000,
"keepaliveCountMax": 3,
"readyTimeout": 20000
}
}
}
}execute_command
Execute a command in the specified shell
Inputs:
shell (string): Shell to use ("powershell", "cmd", or "gitbash")
command (string): Command to execute
workingDir (optional string): Working directory
Returns command output as text, or error message if execution fails
get_command_history
Get the history of executed commands
Input: limit (optional number)
Returns timestamped command history with outputs
ssh_execute
Execute a command on a remote system via SSH
Inputs:
connectionId (string): ID of the SSH connection to use
command (string): Command to execute
Returns command output as text, or error message if execution fails
ssh_disconnect
Disconnect from an SSH server
Input:
connectionId (string): ID of the SSH connection to disconnect
Returns confirmation message
create_ssh_connection
Create a new SSH connection
Inputs:
connectionId (string): ID for the new SSH connection
connectionConfig (object): Connection configuration details including host, port, username, and either password or privateKeyPath
Returns confirmation message
read_ssh_connections
Read all configured SSH connections
Returns a list of all SSH connections from the configuration
update_ssh_connection
Update an existing SSH connection
Inputs:
connectionId (string): ID of the SSH connection to update
connectionConfig (object): New connection configuration details
Returns confirmation message
delete_ssh_connection
Delete an SSH connection
Input:
connectionId (string): ID of the SSH connection to delete
Returns confirmation message
get_current_directory
Get the current working directory of the server
Returns the current working directory path
SSH Connections
URI format: ssh://{connectionId}
Contains connection details with sensitive information masked
One resource for each configured SSH connection
Example: ssh://raspberry-pi shows configuration for the "raspberry-pi" connection
SSH Configuration
URI: ssh://config
Contains overall SSH configuration and all connections (with passwords masked)
Shows settings like defaultTimeout, maxConcurrentSessions, and the list of connections
Current Directory
URI: cli://currentdir
Contains the current working directory of the CLI server
Shows the path where commands will execute by default
CLI Configuration
URI: cli://config
Contains the CLI server configuration (excluding sensitive data)
Shows security settings, shell configurations, and SSH settings
The following security features are hard-coded into the server and cannot be disabled:
Case-insensitive command blocking: All command blocking is case-insensitive (e.g., "DEL.EXE", "del.cmd", etc. are all blocked if "del" is in blockedCommands)
Smart path parsing: The server parses full command paths to prevent bypass attempts (blocking "C:\Windows\System32\rm.exe" if "rm" is blocked)
Command parsing intelligence: False positives are avoided (e.g., "warm_dir" is not blocked just because "rm" is in blockedCommands)
Input validation: All user inputs are validated before execution
Shell process management: Processes are properly terminated after execution or timeout
Sensitive data masking: Passwords are automatically masked in resources (replaced with ********)
These security features are configurable through the config.json file:
Command blocking: Commands specified in blockedCommands array are blocked (default includes dangerous commands like rm, del, format)
Argument blocking: Arguments specified in blockedArguments array are blocked (default includes potentially dangerous flags)
Command injection protection: Prevents command chaining (enabled by default through enableInjectionProtection: true)
Working directory restriction: Limits command execution to specified directories (enabled by default through restrictWorkingDirectory: true)
Command length limit: Restricts maximum command length (default: 2000 characters)
Command timeout: Terminates commands that run too long (default: 30 seconds)
Command logging: Records command history (enabled by default through logCommands: true)
These are not features but important security considerations to be aware of:
Environment access: Commands may have access to environment variables, which could contain sensitive information
File system access: Commands can read/write files within allowed paths - carefully configure allowedPaths to prevent access to sensitive data
This project is licensed under the MIT License - see the LICENSE file for details.
Create a new SSH connection
| Name | Required | Description | Default |
|---|---|---|---|
| connectionId | No | ID of the SSH connection | |
| connectionConfig | No |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations, the description carries full burden but only states the action without behavioral details. It doesn't disclose if this requires authentication, has side effects (e.g., storing credentials), involves rate limits, or what happens on success/failure. For a tool that likely handles sensitive SSH data, this is a significant gap in transparency.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is a single, efficient sentence with zero waste—'Create a new SSH connection' is front-loaded and appropriately sized for its minimal content. It earns its place by stating the core purpose without unnecessary elaboration.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given the complexity (SSH connection creation with nested objects), no annotations, no output schema, and incomplete parameter coverage, the description is inadequate. It doesn't explain what 'create' means operationally, return values, or error conditions, leaving critical gaps for a tool that likely involves network operations and credential handling.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema description coverage is 50% (2 parameters total, with descriptions for nested properties but not top-level ones). The description adds no parameter semantics beyond the schema, which already documents host, port, etc., but doesn't clarify the relationship between 'connectionId' and 'connectionConfig'. Baseline 3 is appropriate as the schema does moderate work.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description 'Create a new SSH connection' clearly states the action (create) and resource (SSH connection), but it's vague about what 'create' entails—does it establish a live connection, store configuration, or both? It doesn't differentiate from siblings like 'update_ssh_connection' or 'ssh_connect' (if present), leaving ambiguity in scope.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
No guidance is provided on when to use this tool versus alternatives. With siblings like 'update_ssh_connection', 'ssh_disconnect', and 'read_ssh_connections', the description lacks context on prerequisites (e.g., whether a connection must not exist) or when to choose creation over updating, offering no usage instructions.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Delete an existing SSH connection
| Name | Required | Description | Default |
|---|---|---|---|
| connectionId | Yes | ID of the SSH connection to delete |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations provided, the description carries full burden for behavioral disclosure. It states this is a deletion operation but doesn't mention whether it's reversible, what permissions are required, what happens to associated resources, or error conditions. For a destructive operation with zero annotation coverage, this is inadequate.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is a single, efficient sentence with zero wasted words. It's appropriately sized for a simple deletion tool and gets straight to the point without unnecessary elaboration.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
For a destructive operation with no annotations and no output schema, the description is insufficient. It doesn't explain what 'delete' entails (permanent removal? soft delete?), what gets returned, or error handling. Given the complexity of SSH connection management and lack of structured data, more context is needed.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema description coverage is 100%, with the single parameter 'connectionId' fully documented in the schema. The description doesn't add any additional meaning about the parameter beyond what the schema already provides, so it meets the baseline for high schema coverage.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the action ('Delete') and target resource ('an existing SSH connection'), providing specific verb+resource pairing. However, it doesn't differentiate this from sibling tools like 'ssh_disconnect' or 'update_ssh_connection', which might have overlapping functionality.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description provides no guidance on when to use this tool versus alternatives like 'ssh_disconnect' or 'update_ssh_connection', nor does it mention prerequisites (e.g., needing an existing connection ID). It simply states what the tool does without context for selection.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Execute a command in the specified shell (powershell, cmd, or gitbash)
Example usage (PowerShell):
{
"shell": "powershell",
"command": "Get-Process | Select-Object -First 5",
"workingDir": "C:\Users\username"
}Example usage (CMD):
{
"shell": "cmd",
"command": "dir /b",
"workingDir": "C:\Projects"
}Example usage (Git Bash):
{
"shell": "gitbash",
"command": "ls -la",
"workingDir": "/c/Users/username"
}| Name | Required | Description | Default |
|---|---|---|---|
| shell | Yes | Shell to use for command execution | |
| command | Yes | Command to execute | |
| workingDir | No | Working directory for command execution (optional) |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations provided, the description carries the full burden of disclosing behavior. It gives examples but does not mention return values, exit codes, error handling, side effects, environment specifics, or security implications. For a command execution tool, this is insufficient.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The opening one-sentence description is concise and front-loaded. The three examples are redundant in structure but serve to illustrate cross-shell syntax. Each sentence earns its place, though the examples could be trimmed to one or two without loss.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
For a tool that executes arbitrary system commands, there is no mention of output format, timeouts, working directory defaults, or safety considerations. The presence of sibling ssh_execute suggests a need to clarify local vs remote execution, which is absent. The examples help but do not fill all gaps.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
The schema already covers 100% of parameters, but the description adds value through concrete examples showing valid shell values, command syntax, and workingDir formatting (e.g., Windows paths vs Git Bash paths). This enhances understanding beyond the schema's terse descriptions.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the tool executes a command in a specified shell (powershell, cmd, or gitbash), which is a specific verb+resource. It is distinguishable from siblings like get_command_history and ssh_execute, though it does not explicitly call out these differences. The mention of the shell options adds specificity.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
No guidance is given on when to use this tool versus alternatives such as ssh_execute (remote execution) or get_command_history. The examples imply usage but do not provide context, prerequisites, or exclusions. This is a clear gap.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Get the history of executed commands
Example usage:
{
"limit": 5
}Example response:
[
{
"command": "Get-Process",
"output": "...",
"timestamp": "2024-03-20T10:30:00Z",
"exitCode": 0
}
]| Name | Required | Description | Default |
|---|---|---|---|
| limit | No | Maximum number of history entries to return (default: 10, max: 1000) |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations are provided, so the description carries the full burden. It adds transparency by showing a detailed example response with fields (command, output, timestamp, exitCode), which clarifies the return format. However, it does not explicitly state whether history is session-scoped, how results are ordered, or that it has no side effects. Given the absence of annotations, there are notable but not critical gaps.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is concise: a single opening sentence plus two compact JSON examples. The purpose is front-loaded, and every element adds value—the example usage demonstrates the parameter, and the example response shows the expected output structure. No redundancy or filler.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
For a simple one-parameter, no-output-schema tool, the description is quite complete: it states the purpose, shows how to invoke it, and provides a representative response. Minor omissions like ordering or session scope are not critical for a list-history tool. The example response effectively substitutes for an output schema, making the tool actionable for an agent.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
The input schema already fully documents the 'limit' parameter with default (10) and max (1000) values, providing 100% coverage. The description's example usage (limit: 5) reinforces the parameter but adds no new semantic meaning beyond the schema. Baseline 3 is appropriate because the schema does the heavy lifting.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description opens with a clear verb+resource statement: 'Get the history of executed commands'. It distinguishes itself from sibling tools like execute_command (execution) and ssh_execute (remote execution) by focusing specifically on retrieval of prior commands. The example response reinforces the purpose.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description provides a clear usage context: to retrieve previously executed commands, with an example showing the optional limit parameter. It does not explicitly state when not to use it, but the purpose is self-evident and no competing history tools exist among siblings. No explicit exclusions are given, but this is clearly a read-only counterpart to the execution siblings.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Get the current working directory
| Name | Required | Description | Default |
|---|---|---|---|
No parameters | |||
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations are provided, and the description does not disclose any behavioral traits beyond the name. There is no mention of side effects, permissions, or return format. The description carries the full burden but adds no extra context.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is a single, clear sentence with no wasted words. It is appropriately sized for a tool with no parameters and a simple purpose.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given the tool has no parameters, no output schema, and a straightforward action, the description is complete enough. It explains what the tool does without needing further details.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
There are zero parameters and schema coverage is 100% (vacuously). The description adds no additional parameter semantics because there are none. Baseline for zero parameters is 4.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the action and resource: 'Get the current working directory'. It is a specific verb-noun pair that distinguishes itself from sibling tools like execute_command or set_current_directory.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
No explicit when-to-use or when-not-to-use guidance is provided. However, the purpose is straightforward and the tool is a simple getter, so implied usage is clear. No alternatives or exclusions are mentioned.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Read all SSH connections
| Name | Required | Description | Default |
|---|---|---|---|
No parameters | |||
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations are provided, so the description carries the full burden of behavioral disclosure. It states 'read all SSH connections,' implying a read-only operation, but doesn't specify if it returns active connections, historical data, permissions required, or any side effects like rate limits. This leaves significant gaps for a tool with zero annotation coverage.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is a single, efficient sentence with no wasted words. It's front-loaded and directly states the tool's purpose, making it highly concise and well-structured for quick understanding.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given the lack of annotations and output schema, the description is incomplete. It doesn't explain what 'SSH connections' entails (e.g., active sessions, configured profiles), the return format, or behavioral traits like safety or limitations. For a read operation with no structured context, this leaves the agent with insufficient information.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
The input schema has 0 parameters with 100% coverage, so no parameter documentation is needed. The description doesn't add parameter details, which is appropriate, earning a baseline score of 4 for adequately handling the lack of parameters.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the verb ('read') and resource ('SSH connections'), making the purpose immediately understandable. However, it doesn't differentiate from sibling tools like 'get_command_history' or 'ssh_execute', which might also involve reading SSH-related data, so it doesn't reach the highest score.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description provides no guidance on when to use this tool versus alternatives. With siblings like 'get_command_history' or 'ssh_execute' that might retrieve related information, there's no indication of context, prerequisites, or exclusions to help the agent choose appropriately.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Disconnect from an SSH server
Example usage:
{
"connectionId": "raspberry-pi"
}Use this to cleanly close SSH connections when they're no longer needed.
| Name | Required | Description | Default |
|---|---|---|---|
| connectionId | Yes | ID of the SSH connection to disconnect |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations, the description must carry the full burden. It mentions 'cleanly close' but doesn't disclose side effects like what happens to associated sessions, idempotency, or error behavior. It's minimally adequate but not rich.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is short and includes a practical JSON example, which helps the agent. It repeats the concept slightly ('Disconnect' and 'close SSH connections'), but overall it's well-structured and efficient.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
For a simple one-parameter tool with no output schema, the description is adequate. It provides an example and clear use case. It doesn't cover errors or edge cases, but given the low complexity, this is not a significant gap.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema coverage is 100% with the parameter described as 'ID of the SSH connection to disconnect'. The example usage adds an illustrative value (raspberry-pi) but doesn't add semantic meaning beyond the schema, so baseline 3 applies.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the verb 'Disconnect' and the resource 'SSH connection/server', which distinguishes it from siblings like ssh_execute and execute_command. The example usage reinforces the exact action.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
It says 'when they're no longer needed', providing clear context for when to use this tool. It doesn't explicitly mention alternatives or when not to use, but the sibling tool names imply the distinction, so a slight deduction is warranted.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Execute a command on a remote host via SSH
Example usage:
{
"connectionId": "raspberry-pi",
"command": "uname -a"
}Configuration required in config.json:
{
"ssh": {
"enabled": true,
"connections": {
"raspberry-pi": {
"host": "raspberrypi.local",
"port": 22,
"username": "pi",
"password": "raspberry"
}
}
}
}| Name | Required | Description | Default |
|---|---|---|---|
| connectionId | Yes | ID of the SSH connection to use | |
| command | Yes | Command to execute |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations are provided, so the description carries the full burden. It mentions that SSH must be enabled in config.json and gives an example connection, which is useful. However, it does not disclose the command execution behavior (e.g., output format, error handling, potential destructive side-effects) or that it uses the specified password for authentication.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is structured and efficient, with a one-sentence purpose followed by illustrative JSON examples. The config block is large but necessary to show the setup requirements. No wasted words.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
With no output schema and no annotations, the description should explain what the tool returns or what 'execute' entails. It does neither, though it does provide configuration context. For a command execution tool, the lack of mention of output or error behavior leaves a moderate gap.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
The input schema descriptions are minimal ('Command to execute', 'ID of the SSH connection to use'). The description adds meaningful value by showing an example usage and the config structure, clarifying that connectionId refers to a key in the 'connections' object and command is a shell string.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states 'Execute a command on a remote host via SSH', using a specific verb and resource. It distinguishes itself from sibling tools like ssh_disconnect (manage connection) and get_command_history (retrieve history).
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description provides a concrete example and configuration requirement, implying when to use the tool (when you have a configured SSH connection and need to run a command). However, it does not explicitly state when to use it over alternatives like execute_command, nor does it provide exclusion criteria.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Update an existing SSH connection
| Name | Required | Description | Default |
|---|---|---|---|
| connectionId | No | ID of the SSH connection to update | |
| connectionConfig | No |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations provided, the description carries the full burden of behavioral disclosure. It states this is an update operation, implying mutation, but doesn't disclose any behavioral traits like permission requirements, whether changes are reversible, what happens to unspecified fields, error conditions, or rate limits. This is inadequate for a mutation tool with zero annotation coverage.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is a single, efficient sentence that gets straight to the point with zero wasted words. It's appropriately sized for a basic tool description and is perfectly front-loaded with the essential information.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
For a mutation tool with no annotations, no output schema, and incomplete parameter documentation (50% schema coverage), the description is insufficient. It doesn't address what the tool returns, error conditions, side effects, or provide enough context about the update operation. The description should do more to compensate for the lack of structured metadata.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema description coverage is 50% (only 'connectionId' has a description in the schema, while 'connectionConfig' and its nested properties lack descriptions). The description adds no parameter semantics beyond what's implied by the tool name - it doesn't explain what fields can be updated, how to structure the config object, or provide examples. The baseline is 3 since the schema covers some parameters, but the description doesn't compensate for the coverage gap.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the action ('Update') and resource ('an existing SSH connection'), making the purpose immediately understandable. It distinguishes from 'create_ssh_connection' by specifying 'existing', but doesn't explicitly differentiate from other sibling tools like 'delete_ssh_connection' beyond the verb difference.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description provides no guidance on when to use this tool versus alternatives. It doesn't mention prerequisites (e.g., needing an existing connection ID), when not to use it, or how it relates to sibling tools like 'create_ssh_connection' or 'delete_ssh_connection' beyond the basic verb difference.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
create_ssh_connectiondelete_ssh_connectionexecute_commandget_command_historyget_current_directoryread_ssh_connectionsssh_disconnectssh_executeupdate_ssh_connectionScored across 9 tools
Most tools have distinct purposes: SSH connection management (create/delete/read/update), SSH operations (execute/disconnect), and local shell operations (execute_command/get_history/get_directory). However, execute_command and ssh_execute both execute commands but in different contexts (local vs remote), which could cause minor confusion if not carefully distinguished by the agent.
All tool names follow a consistent verb_noun pattern with snake_case throughout (e.g., create_ssh_connection, execute_command, get_command_history). The naming is predictable and readable, with no deviations in style or convention.
With 9 tools, the count is well-scoped for a Windows CLI server covering SSH management and command execution. Each tool earns its place by addressing specific operations like connection lifecycle, remote execution, and local shell interactions, without being overly sparse or bloated.
The tool set provides good coverage for SSH connection CRUD (create, read, update, delete) and remote command execution, plus local shell operations. A minor gap exists in local file operations (e.g., read/write files) which could enhance the CLI functionality, but core workflows are adequately supported without dead ends.
The Remote MCP server acts as a standardized bridge between LLM applications (like Claude, ChatGPT, and Cursor) and external services, enabling AI agents to access external tools and resources. Its primary capability is providing a centralized search tool to discover other MCP servers and their respective tools. Unlike local implementations, it runs remotely with OAuth authentication and permission controls for security.
A comprehensive Model Context Protocol (MCP) server that enables AI assistants to interact with yo…
Cloud-hosted MCP server for secure AI access to enterprise data sources via CData Connect AI.
MCP server unifying ERPs, CRMs, APIs and knowledge base for Claude, ChatGPT and Gemini.
