Input/output safety for AI agents: known-pattern injection and obfuscation scan, URL/IP, secrets.
Setup
Choose your client, copy the config, then fill in any required environment variables.
claude mcp add agent-firewall -- npx -y agent-firewall-mcp~/.codex/config.toml
[mcp_servers.agent-firewall]
command = "npx"
args = ["-y", "agent-firewall-mcp"]
# codex mcp add agent-firewall -- npx -y agent-firewall-mcp~/.cursor/mcp.json
{
"mcpServers": {
"agent-firewall": {
"command": "npx",
"args": [
"-y",
"agent-firewall-mcp"
]
}
}
}.vscode/mcp.json
{
"servers": {
"agent-firewall": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"agent-firewall-mcp"
]
}
}
}~/.gemini/settings.json
{
"mcpServers": {
"agent-firewall": {
"command": "npx",
"args": [
"-y",
"agent-firewall-mcp"
]
}
}
}claude_desktop_config.json
{
"mcpServers": {
"agent-firewall": {
"command": "npx",
"args": [
"-y",
"agent-firewall-mcp"
]
}
}
}.trae/mcp.json
{
"mcpServers": {
"agent-firewall": {
"command": "npx",
"args": [
"-y",
"agent-firewall-mcp"
]
}
}
}