Start a new interactive process session. Returns a sessionId for subsequent interactions. Use for long-running processes, REPLs, SSH, or any process requiring stdin/stdout interaction.
AI agents invoke start_process to trigger actions in OODA Computer Control. What it does depends on the arguments the agent supplies, and its effects often reach beyond the immediate call — builds kicked off, notifications sent, workflows started.
This tool allows execution of arbitrary processes and commands via stdin/stdout interaction, including shell commands, SSH sessions, and REPLs. The ability to start interactive processes with stdin/stdout control represents a critical Execute risk—an AI agent could use this to run malicious code, access external systems, exfiltrate data, or compromise system integrity.
From the tool's definition The tool 'start_process' is described as starting 'a new interactive process session' that supports 'long-running processes, REPLs, SSH, or any process requiring stdin/stdout interaction.' This enables arbitrary command execution through interactive shell…
Documented attack patterns abuse exactly the kind of access start_process gives an agent:
PolicyLayer is an MCP gateway — it sits between your AI agents and OODA Computer Control, and nothing reaches the server without passing your rules. This is the rule we recommend for start_process:
{
"version": "1",
"default": "deny",
"tools": {
"start_process": {
"limits": [
{
"counter": "start_process_rate",
"window": "minute",
"max": 10,
"scope": "grant"
}
]
}
}
}start_process stays usable, but rate-capped — a runaway agent can't fire it dozens of times a minute. Everything else on the server is denied unless you say otherwise.
Free to start. No card required.
Start a new interactive process session. Returns a sessionId for subsequent interactions. Use for long-running processes, REPLs, SSH, or any process requiring stdin/stdout interaction. It is categorised as a Execute tool in the OODA Computer Control MCP Server, which means it can trigger actions or run processes. Use rate limits and argument validation.
Register the OODA Computer Control MCP server in PolicyLayer and add a rule for start_process: allow, deny, rate-limit, or require approval. Point your MCP client at the PolicyLayer proxy URL and the rule is enforced on every call, before it reaches OODA Computer Control. Nothing to install.
start_process is a Execute tool with high risk. Execute tools should be rate-limited and have argument validation enabled.
Yes. Add a rate_limit block to the start_process rule in your PolicyLayer policy. For example, setting max: 10 and window: 60 limits the tool to 10 calls per minute. Rate limits are tracked per agent session and reset automatically.
Set action: deny in the PolicyLayer policy for start_process. The AI agent will receive a policy violation error and cannot call the tool. You can also include a reason field to explain why the tool is blocked.
start_process is provided by the OODA Computer Control MCP server (mnehmos/mnehmos.ooda.mcp). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Start from OODA Computer Control, add the rest of your stack, and see everything your agents can call. Then put policy on all of it.
Free to start. No card required.
99 OODA Computer Control tools catalogued and risk-classified — across an index of 43,000+ MCP servers.