
Procm
io.github.coder-kaPublished by GitHub account coder-kav0.0.43Updated Oct 24, 2026
A Model Context Protocol (MCP) server for process management.
Overview
Local MCP server that lets an assistant start, monitor, restart and stop whitelisted processes on your machine and read their output.
- What you can do
- Start processes with a script, arguments, working directory and environment variables, optionally under a friendly name
- Whitelist specific command and working-directory combinations so later starts need no confirmation
- List managed processes and get detailed information about one by ID
- Restart or delete a process, with SIGTERM followed by SIGKILL after 10 seconds
- Read recent stdout and stderr log entries from a process
- When to use it
- Useful when you want an assistant to launch development servers, docker-compose stacks or test watchers in the background, then read their logs and iterate on bugs without you copying output by hand. Processes it created are cleaned up when the session ends.
- Requirements
- Runs as a local stdio process on the desktop; install the npm package procm-mcp and point the client at its built entry file with node. No accounts, API keys or environment variables are declared.
Installation
In SourceWeft
- Open Procm in the dashboard and add it to a workspace.
- Enable the server for the chats that should use its tools.
Desktop only via STDIO. STDIO servers start a local process, so they need the SourceWeft desktop host.
Other MCP clients
Follow the launch instructions in the repository.
README
procm-mcp
A Model Context Protocol (MCP) server for process management.
Supported features
- Secure and automatable process creation
- Cleanup created processes automatically on termination (e.g. exiting claude code)
- Common process management features supported, restarting, deleting, checking status or retreving stdout/stderr of processes
Using these features, LLMs start processes like development servers, docker-compose, or test watchers and check their outputs to fix bugs automatically.
Installation
.mcp.json
Secure process creation
You can permit LLMs to use start-process tool without confirmation, because procm-mcp only allow whitelisted process creations.
LLMs will ask you to use allow-start-process tool to add specific process creation to the whitelist.
Once you allow a process creation, you don't have to confirming it anymore as long as the command and the working directory are the same.
I call it "allow-x pattern", which can balances security and usability in MCP.
Warning: Do not permit LLMs to use allow-start-process without confirmation.That means "Do anything you want to".
Tools
allow-start-processAllow specific processes to be createdscript(required): The script/command to allowargs(optional): Array of argumentscwd(optional): Working directory
start-processStart a new process with specified script and argumentsscript(required): The script/command to executename(optional): A friendly name for the processargs(optional): Array of arguments to pass to the scriptcwd(required): Working directory for the processenvs(optional): Environment variables to set for the process
delete-processStop and remove a process by ID.The default signal is SIGTERM, but SIGKILL(force killing) will be sent after 10 seconds unless the process exits.id(required): The process ID
restart-processRestart an existing process by IDid(required): The process ID
get-process-infoGet detailed information about a processid(required): The process ID
list-processesList all currently managed processes- No parameters required
get-process-stdoutRetrieve stdout logs from a processid(required): The process IDchunkCount(optional): Number of recent log entries to retrieve (default: 10)
get-process-stderrRetrieve stderr logs from a processid(required): The process IDchunkCount(optional): Number of recent log entries to retrieve (default: 10)
License
MIT
Source: README.md at commit 15dcee3
Tools
0Version history
1- v0.0.43LatestSep 16, 2026
Score
40% of the applicable checks met
How complete this listing is, computed from the registry entry, the endpoint and public sources. Not a rating of the server and not a statement that it is safe or working today: reachability is a dated check, and no installs or reviews go into the score.
- Reached in the last 30 days25 pts
Keep the endpoint answering the MCP handshake; SourceWeft probes it every 30 days.
- Auth declared10 pts
Declare the credential headers in the registry entry, or keep the endpoint answering the probe so its auth can be classified.
- Tools listed15 pts
Answer tools/list without a credential, or ship a package that starts in a sandbox without one.
- DescriptionRequired15 pts
- Logo10 pts
- Documentation15 pts
- License10 pts
Declare a license in the repository or the registry entry.