AI agent governance: content scanning, audit logs, policy evaluation, session management.
Setup
Choose your client, copy the config, then fill in any required environment variables.
claude mcp add bulwark --env BULWARK_SESSION_TOKEN='<BULWARK_SESSION_TOKEN>' -- docker run -i --rm -e BULWARK_SESSION_TOKEN ghcr.io/bpolania/bulwark:v0.2.0~/.codex/config.toml
[mcp_servers.bulwark]
command = "docker"
args = ["run", "-i", "--rm", "-e", "BULWARK_SESSION_TOKEN", "ghcr.io/bpolania/bulwark:v0.2.0"]
env = { BULWARK_SESSION_TOKEN = "<BULWARK_SESSION_TOKEN>" }
# codex mcp add bulwark --env BULWARK_SESSION_TOKEN='<BULWARK_SESSION_TOKEN>' -- docker run -i --rm -e BULWARK_SESSION_TOKEN ghcr.io/bpolania/bulwark:v0.2.0~/.cursor/mcp.json
{
"mcpServers": {
"bulwark": {
"command": "docker",
"args": [
"run",
"-i",
"--rm",
"-e",
"BULWARK_SESSION_TOKEN",
"ghcr.io/bpolania/bulwark:v0.2.0"
],
"env": {
"BULWARK_SESSION_TOKEN": "<BULWARK_SESSION_TOKEN>"
}
}
}
}.vscode/mcp.json
{
"servers": {
"bulwark": {
"type": "stdio",
"command": "docker",
"args": [
"run",
"-i",
"--rm",
"-e",
"BULWARK_SESSION_TOKEN",
"ghcr.io/bpolania/bulwark:v0.2.0"
],
"env": {
"BULWARK_SESSION_TOKEN": "<BULWARK_SESSION_TOKEN>"
}
}
}
}~/.gemini/settings.json
{
"mcpServers": {
"bulwark": {
"command": "docker",
"args": [
"run",
"-i",
"--rm",
"-e",
"BULWARK_SESSION_TOKEN",
"ghcr.io/bpolania/bulwark:v0.2.0"
],
"env": {
"BULWARK_SESSION_TOKEN": "<BULWARK_SESSION_TOKEN>"
}
}
}
}claude_desktop_config.json
{
"mcpServers": {
"bulwark": {
"command": "docker",
"args": [
"run",
"-i",
"--rm",
"-e",
"BULWARK_SESSION_TOKEN",
"ghcr.io/bpolania/bulwark:v0.2.0"
],
"env": {
"BULWARK_SESSION_TOKEN": "<BULWARK_SESSION_TOKEN>"
}
}
}
}.trae/mcp.json
{
"mcpServers": {
"bulwark": {
"command": "docker",
"args": [
"run",
"-i",
"--rm",
"-e",
"BULWARK_SESSION_TOKEN",
"ghcr.io/bpolania/bulwark:v0.2.0"
],
"env": {
"BULWARK_SESSION_TOKEN": "<BULWARK_SESSION_TOKEN>"
}
}
}
}Environment variables
BULWARK_SESSION_TOKEN
Replace the placeholders with your own credentials and keep them out of version control.