Triggercmd
TRIGGERcmd is a cloud service that allows you to securely and remotely run commands on your computers.
Verdict
Common use cases
- Deploy code to staging from chat
- Restart hung services during incidents
- Run nightly test suites on demand
- Wake lab machines before demos
- Collect logs from remote servers
Integration
- Vendor
- Triggercmd
- Category
- developer-tools
- Auth
- API_KEY
- Tools
- 3
- Composio slug
triggercmd
Tools
- List TriggerCMD Commands
Tool to retrieve a list of all commands across your computers. use when you need to browse your available commands before triggering one.
- List TriggerCMD Computers
Tool to list all computers associated with your triggercmd account. use after authenticating with your token to retrieve connected machines.
- Trigger Command
Tool to trigger a specified command on a target computer. use when you want to remotely execute a pre-configured command after authentication.
Setup
Setup guide
- 11. Sign up at triggercmd.com and install the Triggercmd agent on each computer you want to control. 2. In the Triggercmd dashboard, define the commands you want to expose (shell scripts, executables, or system actions). 3. Copy your API token from the Triggercmd account settings page. 4. In Switchy, open your workspace settings and navigate to the MCP integrations tab. 5. Select Triggercmd from the catalog and paste your API token when prompted. 6. Click Connect and wait for the green confirmation. 7. Open any Space and type '@Triggercmd list computers' to verify the connection — you should see your registered machines. 8. Type '@Triggercmd list commands' to confirm your pre-configured commands appear. 9. To execute a command, use '@Triggercmd trigger [command name] on [computer name]' and the AI will invoke it remotely.
What teammates see: by default, memories from Triggercmd are scoped to the Space (PROJECT visibility) - you can mark any memory PRIVATE or share it ORG-wide.
Works well with
Top models
Compatibility data appears once enough Spaces have used this MCP together with a given model.
How Switchy teams use it
Starter prompts
List All Machines
@Triggercmd list all computers in my accountOpen in a Space →
Browse Available Commands
@Triggercmd show me all commands I can runOpen in a Space →
Deploy to Staging
@Triggercmd trigger the deploy-staging command on my staging-server machineOpen in a Space →
Restart a Service
@Triggercmd run the restart-nginx command on prod-web-01Open in a Space →
Collect System Logs
@Triggercmd trigger collect-logs on backend-server and summarize the outputOpen in a Space →
Example outputs
Illustrative - representative of the model's voice and quality, not literal recordings.
@Triggercmd show me all the computers connected to my account and their available commands
Found 3 computers in your TriggerCMD account: 'home-desktop' (online, 8 commands including backup-photos, restart-server), 'office-laptop' (offline, 5 commands), and 'media-center' (online, 3 commands including start-plex, shutdown). Each computer shows its current connection status and the pre-configured commands you can trigger remotely.
This read-heavy example demonstrates browsing your remote-execution inventory. The MCP returns only commands you've already configured in TriggerCMD — it cannot create new commands or scripts. Offline computers appear in the list but cannot execute triggers until they reconnect.
@Triggercmd run the backup-photos command on my home-desktop computer right now
Command 'backup-photos' triggered successfully on 'home-desktop'. The TriggerCMD agent on that machine has received the execution request and should begin running the associated script within seconds. Note that this action initiates the command but does not wait for completion or return output from the script itself.
This write example shows the core remote-execution capability. The trigger is fire-and-forget — the MCP confirms the command was sent, but you won't see stdout/stderr or know if the script succeeded unless you've configured separate logging. Requires the target computer to be online and running the TriggerCMD agent.
@Triggercmd which of my computers can restart services, and trigger a restart on whichever one is currently online
Scanning your 3 computers for restart-related commands: 'home-desktop' has 'restart-server' (online), 'office-laptop' has 'restart-docker' (offline). Since 'home-desktop' is online, I've triggered 'restart-server' there. The command has been sent to the TriggerCMD agent and should execute momentarily. The office laptop's restart capability is unavailable until it reconnects.
This synthesis example pairs the MCP's read tools with AI reasoning to filter and act conditionally. Useful for workflows like 'restart whichever dev environment is running' or 'trigger backups on all online machines'. The AI interprets command names heuristically — ensure your TriggerCMD command labels clearly indicate their purpose.
Use-case deep-dives
When TriggerCMD beats SSH for morning boot routines
A 3-person backend team runs Docker stacks on local dev machines that sleep overnight. Every morning standup starts with 'wait, my containers aren't up yet.' TriggerCMD wins here if you've already scripted the startup sequence—one Switchy command fires 'docker-compose up' on the right machine without SSH keys or VPN hassle. The MCP's three tools (list computers, list commands, trigger) are enough for this narrow use-case. Trade-off: if your team needs real-time logs or interactive debugging, you're back to SSH anyway. This MCP shines when the command is fire-and-forget and you want the AI to pick the right machine from your fleet. If your startup routine is more than 5 steps, script it first, then point TriggerCMD at the script.
Triggering deployment approvals from Slack threads
A 6-person product team uses a staging-to-prod pipeline with a manual approval gate—someone has to SSH into the CI runner and run './approve-deploy.sh' after QA sign-off. TriggerCMD replaces that SSH step: the AI in Switchy reads the Slack thread, confirms QA thumbs-up, lists the CI runner in TriggerCMD's computer inventory, and fires the approval command. This works because the gate is a single command with no interactive prompts. The API key auth is fine for internal tooling. Boundary: if your approval logic needs to parse logs or check multiple conditions, TriggerCMD's three-tool scope is too narrow—you'd want a full CI API integration instead. Use this MCP when the human decision is already made and you just need a remote button press.
When TriggerCMD handles post-demo cleanup for sales
A 2-person sales engineering team runs live product demos on a dedicated demo server. After each call, someone manually resets the database and clears test data—a 3-minute SSH chore that happens 4 times a day. TriggerCMD automates this: the AI triggers a pre-configured 'reset-demo-db' command on the demo box right after the calendar event ends. The MCP's simplicity is the win—no need for a full server management platform when you have 5 reset scripts total. Trade-off: if your reset logic changes weekly or needs conditional branching, maintaining the command list in TriggerCMD becomes friction. This MCP is ideal when your remote commands are stable, low-frequency, and don't need output parsing. If you're resetting more than one server, list the computers first so the AI picks the right target.
Frequently asked
What does the TriggerCMD MCP do in Switchy?
It lets your team remotely execute pre-configured shell commands on any computer registered to your TriggerCMD account. You can list available machines, browse commands, and trigger them directly from Switchy's chat interface. Useful for kicking off builds, restarting services, or running scripts on remote dev boxes without SSH.
Do I need admin access to connect TriggerCMD?
You need a TriggerCMD API key, which you generate from your account dashboard. No special admin role required, but whoever connects it controls which commands the team can run. The key grants access to all computers and commands in that TriggerCMD account, so treat it like a production credential.
Can the MCP create new commands or only run existing ones?
It only triggers commands you've already configured in the TriggerCMD agent on each machine. You can't define new commands or edit parameters through Switchy. If you need a new command, add it in the TriggerCMD desktop app first, then it'll appear in the list.
How is this different from SSH or running scripts directly?
TriggerCMD sits between full SSH access and zero control. You pre-approve a whitelist of commands per machine, so teammates can run them without shell access or knowing server IPs. Faster than opening a terminal for routine tasks, safer than handing out SSH keys to everyone.
Who on the team should connect this integration?
Whoever owns the TriggerCMD account and manages the command whitelist. That person's API key becomes the execution identity for the whole workspace. If they leave or revoke the key, all team members lose access until someone reconnects with a valid credential.