Start an agent in a specific tmux target
AI agents invoke start_agent to trigger actions in Agent Collaboration MCP Server. 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 executes external operations by spawning agents in tmux sessions. While launching a process itself is reversible (agents can be stopped), the effects of what those agents do are not controlled by this tool and depend entirely on the agent's implementation and arguments.
From the tool's definition start_agent: 'Start an agent in a specific tmux target' — this launches new agents in tmux sessions, which are external processes that can execute arbitrary code and operations.
Documented attack patterns abuse exactly the kind of access start_agent gives an agent:
PolicyLayer is an MCP gateway — it sits between your AI agents and Agent Collaboration MCP Server, and nothing reaches the server without passing your rules. This is the rule we recommend for start_agent:
{
"version": "1",
"default": "deny",
"tools": {
"start_agent": {
"limits": [
{
"counter": "start_agent_rate",
"window": "minute",
"max": 10,
"scope": "grant"
}
]
}
}
}start_agent 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 an agent in a specific tmux target. It is categorised as a Execute tool in the Agent Collaboration MCP Server MCP Server, which means it can trigger actions or run processes. Use rate limits and argument validation.
Register the Agent Collaboration MCP Server MCP server in PolicyLayer and add a rule for start_agent: 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 Agent Collaboration MCP Server. Nothing to install.
start_agent 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_agent 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_agent. 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_agent is provided by the Agent Collaboration MCP Server MCP server (nishimoto265/agent_collaboration_mcp). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Start from Agent Collaboration MCP Server, 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.
6 Agent Collaboration MCP Server tools catalogued and risk-classified — across an index of 43,000+ MCP servers.