claude mcp add arvisn4-agent-guardrail -- uvx agent-guardrail
{
"mcpServers": {
"arvisn4-agent-guardrail": {
"command": "uvx",
"args": ["agent-guardrail"]
}
}
}
{
"mcpServers": {
"arvisn4-agent-guardrail": {
"command": "uvx",
"args": ["agent-guardrail"]
}
}
}
{
"mcpServers": {
"arvisn4-agent-guardrail": {
"command": "uvx",
"args": ["agent-guardrail"]
}
}
}
About This MCP Server
The agent-guardrail server is a developer tool that provides action-level governance for AI agents, enabling control over the actions they perform rather than just their output. It connects to AI coding assistants through standard input/output interfaces, allowing developers to set boundaries on agent behavior. Developers would use this tool when they need to ensure that AI agents are only performing authorized actions, particularly in applications where security and compliance are critical.
Action-level governance for AI agents -- control what they DO, not what they SAY. This is a Model Context Protocol (MCP) server that extends AI assistants like Claude with developer tools capabilities via the stdio transport.
Package
agent-guardrail on pypi
HOW TO USE
Select your AI client above to get the install command. This MCP server uses the stdio transport and is available on pypi.
What tools does it provide?
The Agent Guardrail MCP Server server extends your AI assistant with developer tools capabilities. Once installed, your AI can use its tools automatically.
