The Mac Shell MCP Server enables secure execution of macOS terminal commands through the Model Context Protocol with comprehensive security controls and approval workflows.
Command Execution: Execute macOS terminal commands using ZSH shell through the execute_command tool with optional arguments, subject to whitelist and security level checks.
Whitelist Management:
View whitelisted commands and their security levels with get_whitelist
Add new commands using add_to_whitelist with specified security level and optional description
Update security levels of existing commands via update_security_level
Remove commands from the whitelist using remove_from_whitelist
Approval Workflow:
View pending commands awaiting approval with get_pending_commands
Approve or deny pending commands using approve_command and deny_command (with optional reason)
Security Levels:
Safe: Execute immediately without approval (e.g., ls, pwd, cat, echo)
Requires approval: Held in queue until explicitly approved (e.g., mv, cp, mkdir)
Forbidden: Explicitly blocked from execution (e.g., rm, sudo)
Security Features:
Uses execFile to prevent shell injection attacks
Validates arguments against allowed patterns
Pre-configured whitelist with common safe commands included by default
Commands run with the permissions of the user running the server
Compatible with Roo Code and Claude Desktop via local installation or npx
Click on "Install 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., "@Mac Shell MCP Server list all files in my Downloads folder"
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.
An MCP server that lets an AI client run a small, fixed set of read-only shell commands, confined to directories you nominate.
It is built around one idea: the agent cannot widen its own authority. There is no tool that edits the policy, no approval queue the agent can drain, and no shell to interpret its arguments.
Upgrading from 1.x? 2.0.0 is a breaking change. Seven tools were removed and the authorization model was replaced. See Migrating from 1.x. 1.x contained two vulnerabilities reported by four independent researchers; see the security advisories.
Tool | What it is for |
| Run a permitted command inside your configured directories |
| Run one outside them — requires a client that can ask a human |
| Chain read-only commands, stdout to stdin, without a shell |
| Ask what is permitted, rather than guessing and being refused |
| Turn observed usage into config you apply |
Related MCP server: Terminal MCP Server
ls pwd echo cat head tail wc grep
Each is restricted to an allowlist of flags. None of them can write a file, execute another program, or read configuration from the environment.
find, git, rm and every interpreter are absent by default. You can add any of them, and should know why they are not there:
find -fprintf writes attacker-chosen content to an attacker-chosen path, while find is naturally classified read-only
git executes commands from a repository's own .git/config, and no environment variable disables that
interpreters (python, perl, awk, node, osascript, …) take inline code as an argument
npm install -g @the_cfdude/mac-shell-mcpThen create a policy:
mkdir -p ~/.mac-shell-mcp
cp .mac-shell-mcp.sample.json ~/.mac-shell-mcp/config.json
$EDITOR ~/.mac-shell-mcp/config.json # set allowedRoots to your project
chmod 444 ~/.mac-shell-mcp/config.json # defence in depth; see SecurityAdd to your MCP client:
{
"mcpServers": {
"mac-shell": { "command": "npx", "args": ["-y", "@the_cfdude/mac-shell-mcp"] }
}
}A fresh install with no roots configured refuses everything — deliberately. Each refusal names the permitted commands and where configuration lives, so the first failure tells you what to do.
Policy is found in this order, first match wins, never merged:
$MAC_SHELL_MCP_CONFIG — an explicit path
~/.mac-shell-mcp/config.json
Host environment (MAC_SHELL_ROOTS, MAC_SHELL_COMMANDS)
Built-in defaults
Policy is never read from the working directory. A cloned repository must not be able to supply the policy governing the agent that opens it.
{
"allowedRoots": ["/Users/you/Projects/my-project"],
"programDirectories": ["/usr/bin", "/bin", "/usr/sbin", "/sbin"],
"commands": {
"grep": {
"program": "/usr/bin/grep",
"effect": "read",
"allowedArgs": ["-i", "-n", "-r", "-l", "-E", "-F"],
"permission": "allow",
},
},
}permission is allow, ask, or deny. ask requires a client offering MCP elicitation — the only capability that means a human can be asked. Where it is absent, ask becomes deny, never allow.
Authorization is effect × scope, computed per call. grep inside your project is free; the same grep against ~/.aws is not. The command name alone never decides.
No shell. execFile is never given a shell option, so arguments reach the OS uninterpreted. Nothing evaluates ;, $(), or backticks — which is also why filenames with spaces, parentheses and brackets work normally.
The program is authorized, not just the arguments. Commands resolve to an absolute path from your program directories, are matched by that resolved path rather than by basename, and are refused if they resolve inside one of your roots. Roots hold data, never code.
The environment is constructed, not inherited, so a variable like RIPGREP_CONFIG_PATH cannot smuggle in a helper program that never appears in the argument list.
Allowlists, never denylists. A denylist is a list of the flags somebody thought of.
Scope includes the working directory, so a command with no path argument is judged by where it runs rather than passing because it named nothing.
The policy file, its parent directories, your program directories, and the whole audit log directory are protected locations — the server refuses to modify any of them, judged by the operation a request performs rather than by a command's declared effect. This is enforced in code, keyed on filesystem identity rather than path strings, and cannot be switched off from the policy file.
chmod 444 on your config is worth doing as a second, independent layer. It is not the mechanism: replacing a file needs write permission on its directory, not the file.
Every request, permitted or refused, is recorded to an append-only audit log.
Known limits, stated rather than implied — see SECURITY.md:
Path detection is heuristic, and fails closed
Everything inside a root is freely readable by one recursive search
ask is only as strong as the host's approval prompt
Flag meanings vary between implementations of the same command name
Report vulnerabilities via private advisory or security@onvex.ai.
Removed: add_to_whitelist, update_security_level, remove_from_whitelist, approve_command, deny_command, get_pending_commands, get_whitelist.
The first three let any client promote a forbidden command to safe. The next three formed an approval workflow the requesting agent could drain by itself. get_whitelist is replaced by get_policy.
Runtime whitelist edits become a config file you edit and the server reads at startup. rm was FORBIDDEN in 1.x and remains unavailable, so nothing that worked before stops working.
npm install && npm run build
npm test # run it twice; a suite that only passes once is not passing
npm run lint && npm run format:checkDesign and specifications: openspec/changes/mac-shell-mcp-2-security-redesign/.
MIT — see LICENSE.
Add a command to the whitelist
| Name | Required | Description | Default |
|---|---|---|---|
| command | Yes | The command to whitelist | |
| securityLevel | Yes | Security level for the command | |
| description | No | Description of the command |
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 the action ('Add') but doesn't explain what happens upon invocation—e.g., whether it's a mutation, requires permissions, has side effects like notifications, or returns confirmation. This leaves significant gaps for a tool that modifies a security list.
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—it directly states the tool's purpose without unnecessary words. It's appropriately sized and front-loaded, making it easy to parse quickly.
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 of modifying a security whitelist, no annotations, and no output schema, the description is incomplete. It doesn't cover behavioral aspects like mutation effects, error conditions, or return values, which are crucial for safe tool invocation in this context.
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%, so the schema already documents all three parameters (command, securityLevel, description) with details like enum values for securityLevel. The description adds no additional meaning beyond the schema, resulting in the baseline score of 3.
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 ('Add') and resource ('command to the whitelist'), making the purpose immediately understandable. However, it doesn't explicitly differentiate from sibling tools like 'update_security_level' or 'remove_from_whitelist', which prevents a perfect 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?
No guidance is provided about when to use this tool versus alternatives like 'update_security_level' or 'approve_command'. The description lacks context about prerequisites, such as whether the command must be pending or already exist, leaving usage unclear.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Approve a pending command
| Name | Required | Description | Default |
|---|---|---|---|
| commandId | Yes | ID of the command to approve |
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. 'Approve' implies a mutation that changes state, but the description doesn't clarify what happens after approval (e.g., does it trigger execution, log the action, or require additional steps?), potential side effects, permission requirements, or error conditions. This leaves significant gaps for a tool that likely involves security-sensitive operations.
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, direct sentence with no wasted words. It front-loads the core action and resource efficiently, making it easy to parse and understand at a glance.
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, and the tool's likely role in a security/command workflow (inferred from sibling tools), the description is insufficient. It doesn't explain the outcome of approval, how it interacts with other tools (e.g., 'execute_command'), or any system constraints, leaving the agent with incomplete context for safe and effective use.
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 'commandId' clearly documented in the schema. The description adds no additional parameter semantics beyond implying that 'commandId' refers to a pending command, which is already inferred from the tool's purpose. This 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 ('approve') and the target resource ('a pending command'), which is specific and unambiguous. However, it doesn't differentiate this tool from its sibling 'deny_command' beyond the opposite action, nor does it explain what 'approve' entails in this context versus simply executing or modifying the command.
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 'deny_command' or 'execute_command'. It doesn't specify prerequisites (e.g., that the command must be in a pending state) or contextual cues for selection, leaving the agent to infer usage from the tool name alone.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Deny a pending command
| Name | Required | Description | Default |
|---|---|---|---|
| commandId | Yes | ID of the command to deny | |
| reason | No | Reason for denial |
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 the action ('deny') but doesn't explain what 'deny' entails—whether it's reversible, if it requires specific permissions, what the outcome is, or if there are side effects. For a mutation tool with zero annotation coverage, this leaves significant gaps in understanding the tool's behavior.
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 front-loaded with the core action and target, making it easy to parse quickly. Every word earns its place by conveying essential information without redundancy.
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 that this is a mutation tool with no annotations and no output schema, the description is incomplete. It doesn't explain the tool's behavior, outcomes, or error conditions. The agent lacks context on what 'deny' means operationally, making it difficult to use correctly without additional 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 100% description coverage, with clear documentation for both parameters ('commandId' and 'reason'). The description doesn't add any meaning beyond what the schema provides, such as format examples or constraints. Since schema coverage is high, the baseline score of 3 is appropriate, as the schema handles parameter semantics adequately.
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 ('deny') and the target ('a pending command'), providing a specific verb+resource combination. However, it doesn't differentiate from sibling tools like 'approve_command' or explain what 'deny' means in this context beyond the basic 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?
The description provides no guidance on when to use this tool versus alternatives like 'approve_command' or 'execute_command'. It doesn't mention prerequisites (e.g., needing a pending command ID) or context about what happens after denial. The agent must infer usage from the tool name alone.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Execute a shell command on macOS
| Name | Required | Description | Default |
|---|---|---|---|
| command | Yes | The command to execute | |
| args | No | Command arguments |
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 what the tool does but doesn't describe critical behavioral aspects: whether execution is immediate or requires approval, what permissions are needed, whether commands run with elevated privileges, what happens on failure, or what output format to expect. For a potentially dangerous shell execution tool, this is inadequate 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 that communicates the core functionality without unnecessary words. It's appropriately sized for a tool with two parameters and gets straight to the point. Every word earns its place, making it easy to parse quickly.
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 of shell command execution (potentially destructive, security-sensitive) and the absence of both annotations and output schema, the description is insufficiently complete. It doesn't address safety considerations, error handling, output format, or how this interacts with the security-focused sibling tools. For a tool that could modify system state, more contextual information 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 both parameters ('command' and 'args') clearly documented in the schema. The description doesn't add any parameter-specific information beyond what the schema already provides. According to scoring rules, when schema coverage is high (>80%), the baseline is 3 even without additional param details in the description.
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 ('Execute') and target ('a shell command on macOS'), making the purpose immediately understandable. It distinguishes from siblings by focusing on command execution rather than security management or querying. However, it doesn't specify whether this executes immediately or requires approval, which could be more precise.
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 about when to use this tool versus alternatives like 'approve_command' or 'deny_command'. The description doesn't mention prerequisites, security implications, or whether this bypasses approval workflows. With siblings focused on security controls, this omission leaves significant ambiguity about appropriate usage contexts.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Get the list of commands pending approval
| Name | Required | Description | Default |
|---|---|---|---|
No parameters | |||
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 read operation ('Get'), implying it's non-destructive, but doesn't cover other aspects like permissions needed, rate limits, response format, or whether it returns all pending commands or a filtered subset. The description is minimal and lacks behavioral 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's front-loaded with the core purpose and efficiently communicates the tool's function without unnecessary elaboration, making it easy for an agent to parse quickly.
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 annotations, no output schema, and a simple input schema with no parameters, the description is minimal. While it states the purpose, it lacks context about what 'pending approval' means, how commands are structured, or what the return value looks like. For a tool in a security/command approval context, more completeness would be helpful.
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 tool has 0 parameters with 100% schema description coverage, so no parameter documentation is needed. The description doesn't add parameter details, which is appropriate here. A baseline score of 4 is given since the schema fully covers the absence 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 action ('Get') and resource ('list of commands pending approval'), making the purpose immediately understandable. It doesn't explicitly distinguish from sibling tools like 'approve_command' or 'deny_command', but the verb 'Get' implies a read operation rather than an action on the commands themselves.
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 'get_whitelist' or 'execute_command'. It doesn't mention prerequisites, context for pending commands, or any exclusions, leaving the agent to infer usage from the tool name alone.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Get the list of whitelisted commands
| 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 'Get the list' which implies a read-only operation, but doesn't specify whether this requires authentication, returns paginated results, includes metadata, or has rate limits. For a tool with zero annotation coverage, this leaves significant behavioral 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 a single, clear sentence with no wasted words. It's front-loaded with the core purpose ('Get the list of whitelisted commands') and doesn't include unnecessary elaboration. Every word earns its place.
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 and no output schema, the description is minimally adequate but lacks context about the return format (e.g., structure of the whitelist, data types) and behavioral traits. It meets the basic requirement for a simple read operation but doesn't provide enough information for confident use without additional context.
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 tool has 0 parameters with 100% schema description coverage, so the schema fully documents the absence of inputs. The description doesn't need to add parameter details, and it appropriately doesn't mention any. Baseline 4 is correct for zero-parameter tools where the schema already covers everything.
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 ('Get') and the resource ('list of whitelisted commands'), making the purpose immediately understandable. It doesn't explicitly differentiate from siblings like 'get_pending_commands' or 'add_to_whitelist', but the verb+resource combination is specific enough to infer basic distinction.
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, timing, or contrast with sibling tools like 'get_pending_commands' (which might retrieve unapproved commands) or 'add_to_whitelist' (which modifies the list). Usage is implied by the name but not explicitly stated.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Remove a command from the whitelist
| Name | Required | Description | Default |
|---|---|---|---|
| command | Yes | The command to remove from whitelist |
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 the tool performs a removal action, implying mutation, but doesn't describe permissions needed, whether the change is reversible, error conditions, or what happens if the command isn't in the whitelist. This leaves significant gaps for a mutation tool.
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 states the core function without any wasted words. It's appropriately sized for a simple tool with one parameter and gets straight to the point.
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 and no output schema, the description is insufficient. It doesn't explain what happens after removal (e.g., confirmation message, error responses), doesn't mention security implications, and provides no context about the whitelist system it operates within.
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%, so the schema already documents the single 'command' parameter adequately. The description doesn't add any additional meaning about parameter format, constraints, or examples beyond what the schema provides, meeting the baseline for high 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 ('remove') and target resource ('a command from the whitelist'), making the purpose immediately understandable. However, it doesn't differentiate from sibling tools like 'deny_command' or 'update_security_level' which might have overlapping security functions, preventing a perfect 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?
No guidance is provided on when to use this tool versus alternatives like 'deny_command' or 'update_security_level'. The description only states what it does, not when it's appropriate or what prerequisites might exist (e.g., whether the command must already be whitelisted).
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Update the security level of a whitelisted command
| Name | Required | Description | Default |
|---|---|---|---|
| command | Yes | The command to update | |
| securityLevel | Yes | New security level for the command |
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. While 'Update' implies a mutation, it doesn't specify required permissions, whether changes are reversible, error conditions (e.g., if command isn't whitelisted), or side effects. This leaves significant gaps for a security-related mutation tool.
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, direct sentence with zero wasted words. It front-loads the core action and target efficiently, making it easy to parse while conveying essential 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?
For a mutation tool with no annotations and no output schema, the description is insufficient. It lacks critical context: what 'security level' means operationally, how changes affect command execution, error handling, or response format. Given the security-sensitive nature and sibling tools, more completeness 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%, so parameters are fully documented in the schema. The description adds no additional parameter context beyond implying 'command' refers to a whitelisted one. This meets the baseline for high schema coverage but doesn't enhance understanding of parameter usage or constraints.
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 the target ('security level of a whitelisted command'), making the purpose immediately understandable. However, it doesn't explicitly differentiate this tool from siblings like 'add_to_whitelist' or 'remove_from_whitelist', which also modify command permissions but in different ways.
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., the command must already be whitelisted), contrast with 'approve_command'/'deny_command' for pending requests, or explain when changing security levels is appropriate versus adding/removing from the whitelist.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Each tool has a clearly distinct purpose with no ambiguity. The tools cover separate aspects of command management: whitelist operations (add/remove/get), approval workflow (approve/deny/get_pending), execution (execute_command), and security configuration (update_security_level). There is no overlap in functionality.
All tools follow a consistent verb_noun pattern with snake_case throughout. The naming is predictable and readable, using clear action verbs like 'add_to', 'remove_from', 'get', 'execute', 'approve', 'deny', and 'update' paired with appropriate nouns.
With 8 tools, this server is well-scoped for its macOS shell security management purpose. Each tool earns its place by covering essential operations for command whitelisting, approval workflows, execution, and security configuration without being excessive or insufficient.
The tool set provides complete coverage for the macOS shell security domain. It includes full CRUD operations for the whitelist (add, remove, get), comprehensive approval workflow management (approve, deny, get_pending), command execution, and security level updates, leaving no obvious gaps for agents to work with.
Unclaimed servers have limited discoverability.
If you are the server author, to access and configure the admin panel.
Hosted MCP server connecting claude.ai, ChatGPT and other AI apps to your own computer
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 paid remote MCP for CLI tool MCP, built to return verdicts, receipts, usage logs, and audit-ready
An MCP server for Arcjet - the runtime security platform that ships with your AI code.
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/cfdude/mac-shell-mcp'
If you have feedback or need assistance with the MCP directory API, please join our Discord server
