MCP, MCP server
Switchboard
Switchboard, an MCP server listed in the PlugVet directory.
Compatibility: Not stated.
Grade D
Two or more Medium findings, or a High finding. Only try it in a sandbox.
Not a guarantee. A hygiene check of the source at the commit shown, not a security audit.
Rescanned at commit 0c436ecbeccc, unchanged since the check below.
INSTALL_REMOTE_SCRIPT NETWORK_HOST_FROM_INPUT SECRETS_PATH_CREDENTIAL
Why this grade
- INSTALL_REMOTE_SCRIPT High
Fetches a script from the network and runs it (a pipe to a shell, or an install hook with a URL).
server.js, line 18
// `curl http://my-host:PORT/install.sh | sh` self-targeting - no hardcoded IPs. - NETWORK_HOST_FROM_INPUT High
A file takes a URL or endpoint as input and has no fixed host list.
bus.js, line 415
.then(() => fetch(url, { method: "POST", headers: { "content-type": "application/json" }, body: JSON.stringify(body) })) - SECRETS_PATH_CREDENTIAL Medium
Mentions a credential store path (~/.ssh, id_rsa, .npmrc, .aws/credentials, credentials.json, .netrc).
channel/run-channel-session.py, line 23
# dump ~/.claude/.credentials.json or ~/.switchboard/config.json straight to the reply tool.
Any High, or two or more Medium findings, give D. One Medium gives C. Low findings or a security Note give B. Each code names a finding behind the letter.
If you think this grade is wrong, email [email protected] with the entry link and your evidence. We aim to acknowledge within two working days and decide within seven. If we confirm an error, we aim to correct it within 24 hours of confirming it. These are aims, not guarantees.
Grades are a hygiene check of the source at the commit shown, on the date checked, not a security audit or a guarantee. The free directory grades are automated. Before release, a PlugVet engineer reviews every paid report and any grade we hold back, including all F grades, against the source code and the evidence we quote.
What was checked
Plain notes from the hygiene checks.
Permissions
Note
requests shell access. The manifest names this permission.
PERMISSION_SHELL_DECLARED
channel/run-channel-session.py, line 47
p = subprocess.Popen(
daemon/claude-agent-daemon.py, line 203
result = subprocess.run(
waker/switchboard_waker.py, line 77
proc = subprocess.run(
Show more
waker/switchboard_waker.py, line 113
out = subprocess.run([cfg["claude_bin"], "agents", "--json"], capture_output=True, text=True,
waker/switchboard_waker.py, line 168
proc = subprocess.run(cmd, capture_output=True, text=True, timeout=180, stdin=subprocess.DEVNULL,
windows/windows-daemon.py, line 110
# `claude.cmd`. subprocess.run with a list doesn't find .cmd without shell=True
windows/windows-daemon.py, line 309
subprocess.run(
windows/windows-daemon.py, line 335
result = subprocess.run(
Install and update
Serious
install fetches a script from the network.
INSTALL_REMOTE_SCRIPT
README.md, line 63
$ curl -fsSL http://your-host:3107/install.sh | sh -s -- --agent-id myagent --token <token>
server.js, line 18
// `curl http://my-host:PORT/install.sh | sh` self-targeting - no hardcoded IPs.
tools.js, line 198
: `curl -fsSL ${b}/install.sh | sh -s -- --agent-id ${agent_id} --token ${tok}`;
Show more
deploy/quickstart.ps1, line 60
Write-Host " curl -fsSL $base/install.sh | sh -s -- --agent-id myagent --token $token"
deploy/quickstart.sh, line 63
echo " curl -fsSL ${BASE}/install.sh | sh -s -- --agent-id myagent --token ${TOKEN}"
docs/connect-agents.md, line 13
$ curl -fsSL http://your-host:3107/install.sh | sh -s -- --agent-id myagent --token <token>
docs/hooks-and-daemon.md, line 46
$ curl -fsSL http://your-host:3107/install.sh | sh -s -- \
install/install.sh, line 6
# curl -fsSL http://my-switchboard:3107/install.sh | sh
install/install.sh, line 7
# curl -fsSL http://my-switchboard:3107/install.sh | sh -s -- --agent-id billy --with-daemon
Network
Serious
accepts a host from input, with no fixed list in this file.
NETWORK_HOST_FROM_INPUT
bus.js, line 415
.then(() => fetch(url, { method: "POST", headers: { "content-type": "application/json" }, body: JSON.stringify(body) }))
channel/switchboard-channel.mjs, line 54
const r = await fetch(`${BASE}/mcp`, {
daemon/claude-agent-daemon.py, line 131
resp = requests.post(CFG["url"], json=payload, headers=headers, timeout=http_timeout)
Show more
deploy/quickstart.sh, line 45
until curl -sf "http://localhost:${PORT}/healthz" >/dev/null 2>&1; do
hooks/agentbus-digest.mjs, line 25
const r = await fetch(`${cfg.base}/activity?limit=20`, {
hooks/agentbus-publish.mjs, line 34
await fetch(`${cfg.base}/status`, {
hooks/switchboard-digest.mjs, line 48
const r = await fetch(`${cfg.base}/sync`, {
hooks/switchboard-digest.mjs, line 59
const r2 = await fetch(`${cfg.base}/activity?limit=20`, {
hooks/switchboard-publish.mjs, line 56
const r = await fetch(`${cfg.base}/sync`, {
windows/windows-daemon.py, line 170
with urllib.request.urlopen(req, timeout=http_timeout) as resp:
Secrets
Concern
reads a credential or environment file.
SECRETS_ENV_READ SECRETS_PATH_CREDENTIAL SECRETS_PATH_DOTENV
server.js, line 64
if (process.env.SWITCHBOARD_MCP_TOKEN) return process.env.SWITCHBOARD_MCP_TOKEN;
server.js, line 67
const tokenFile = process.env.SWITCHBOARD_TOKEN_FILE || join(dirname(dbPath), "token");
channel/run-channel-session.py, line 23
# dump ~/.claude/.credentials.json or ~/.switchboard/config.json straight to the reply tool.
Show more
channel/switchboard-channel.mjs, line 28
const TOKEN = process.env.SWITCHBOARD_MCP_TOKEN || cfg.token;
channel/switchboard-channel.mjs, line 89
"~/.claude/.credentials.json), and never run destructive or system-modifying actions on the " +
daemon/claude-agent-daemon.py, line 60
token = os.environ.get("SWITCHBOARD_MCP_TOKEN") or cfg.get("token")
deploy/quickstart.ps1, line 17
if (-not $token -and (Test-Path .env)) {
deploy/quickstart.ps1, line 18
$line = Select-String -Path .env -Pattern '^SWITCHBOARD_MCP_TOKEN=(.+)$' | Select-Object -First 1
deploy/quickstart.ps1, line 25
if (-not (Test-Path .env)) { if (Test-Path .env.example) { Copy-Item .env.example .env } else { New-Item -ItemType File .env | Out-Null } }
deploy/quickstart.ps1, line 26
(Get-Content .env | Where-Object { $_ -notmatch '^\s*#?\s*SWITCHBOARD_MCP_TOKEN=' }) | Set-Content .env
deploy/quickstart.ps1, line 27
Add-Content .env "SWITCHBOARD_MCP_TOKEN=$token"
deploy/quickstart.ps1, line 28
Write-Host '-> generated a token and wrote it to .env'
install/install.mjs, line 50
const token = args["token"] || process.env.SWITCHBOARD_MCP_TOKEN;
windows/windows-daemon.py, line 81
token = os.environ.get("SWITCHBOARD_MCP_TOKEN") or cfg.get("token") # pragma: allowlist secret
Agent instructions
Pass
No instruction to ignore earlier instructions or to hide a step.
Obfuscation
Pass
No encoded or minified payload found.
Licence
Pass
A licence file is present (LICENSE).
Maintainer activity
Pass
The latest commit is within the last 365 days.
Note
Hygiene check of commit 0c436ecbeccc. requests shell access. The manifest names this permission. install fetches a script from the network. accepts a host from input, with no fixed list in this file. reads a credential or environment file.
Date and source
Date checked:
github.com
Commit graded: 0c436ecbeccce43c2d09eb7046f14fe383d853ac
This link goes to the original host. PlugVet does not host or check what you download from there, and installing it is your own decision and risk.