MCP servers expose tools with no information about what they actually do at runtime. mcpsafetywarden sits between your agent and any MCP server, profiling tool behavior, blocking destructive calls, and running active security audits before you trust them in a workflow.
Setup
Choose your client, copy the config, then fill in any required environment variables.
claude mcp add safety-warden -- uvx mcpsafetywarden~/.codex/config.toml
[mcp_servers.safety-warden]
command = "uvx"
args = ["mcpsafetywarden"]
# codex mcp add safety-warden -- uvx mcpsafetywarden~/.cursor/mcp.json
{
"mcpServers": {
"safety-warden": {
"command": "uvx",
"args": [
"mcpsafetywarden"
]
}
}
}.vscode/mcp.json
{
"servers": {
"safety-warden": {
"type": "stdio",
"command": "uvx",
"args": [
"mcpsafetywarden"
]
}
}
}~/.gemini/settings.json
{
"mcpServers": {
"safety-warden": {
"command": "uvx",
"args": [
"mcpsafetywarden"
]
}
}
}claude_desktop_config.json
{
"mcpServers": {
"safety-warden": {
"command": "uvx",
"args": [
"mcpsafetywarden"
]
}
}
}.trae/mcp.json
{
"mcpServers": {
"safety-warden": {
"command": "uvx",
"args": [
"mcpsafetywarden"
]
}
}
}