Stop current active debug session and terminate OpenOCD/GDB processes.
AI agents invoke debug_stop to trigger actions in Openocd. 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 a command that terminates running processes and debug sessions. While not destructive to persistent data (it does not delete files or overwrite memory), it is an Execute category action because it runs an external operation (process termination) with effects that depend on the current system state.
From the tool's definition The tool performs 'Stop current active debug session and terminate OpenOCD/GDB processes.' This directly triggers termination of external processes and embedded system debugging sessions, which are external operations whose effects depend on which debug…
Documented attack patterns abuse exactly the kind of access debug_stop gives an agent:
PolicyLayer is an MCP gateway — it sits between your AI agents and Openocd, and nothing reaches the server without passing your rules. This is the rule we recommend for debug_stop:
{
"version": "1",
"default": "deny",
"tools": {
"debug_stop": {
"limits": [
{
"counter": "debug_stop_rate",
"window": "minute",
"max": 10,
"scope": "grant"
}
]
}
}
}debug_stop 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.
Stop current active debug session and terminate OpenOCD/GDB processes. It is categorised as a Execute tool in the Openocd MCP Server, which means it can trigger actions or run processes. Use rate limits and argument validation.
Register the Openocd MCP server in PolicyLayer and add a rule for debug_stop: 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 Openocd. Nothing to install.
debug_stop 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 debug_stop 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 debug_stop. 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.
debug_stop is provided by the Openocd MCP server (luiox/openocd-mcp). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Start from Openocd, 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.
14 Openocd tools catalogued and risk-classified — across an index of 43,000+ MCP servers.