AI agents invoke gemini_switch_model to trigger actions in Gemini Skill. 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 switches the active model in the Gemini web interface, which is an external operation that changes application state via browser automation (CDP). It's not a simple data read or write, but an action that modifies the operational context of the browser-controlled interface. Severity is medium as misuse could cause unintended model switches affecting subsequent operations.
From the tool's definition 切换 Gemini 模型（pro / flash / flash-lite）— 'switch model' triggers an external operation on the Gemini web interface via CDP, changing the active model state
Documented attack patterns abuse exactly the kind of access gemini_switch_model gives an agent:
PolicyLayer is an MCP gateway — it sits between your AI agents and Gemini Skill, and nothing reaches the server without passing your rules. This is the rule we recommend for gemini_switch_model:
{
"version": "1",
"default": "deny",
"tools": {
"gemini_switch_model": {
"limits": [
{
"counter": "gemini_switch_model_rate",
"window": "minute",
"max": 10,
"scope": "grant"
}
]
}
}
}gemini_switch_model 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.
切换 Gemini 模型（pro / flash / flash-lite）. It is categorised as a Execute tool in the Gemini Skill MCP Server, which means it can trigger actions or run processes. Use rate limits and argument validation.
Register the Gemini Skill MCP server in PolicyLayer and add a rule for gemini_switch_model: 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 Gemini Skill. Nothing to install.
gemini_switch_model 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 gemini_switch_model 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 gemini_switch_model. 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.
gemini_switch_model is provided by the Gemini Skill MCP server (wjz-p/gemini-skill). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Deterministic rules across all 21 Gemini Skill tools. Per-identity grants. Full audit log. Live in minutes. Nothing to install.
Free to start. No card required.
21 Gemini Skill tools catalogued and risk-classified — across an index of 42,500+ MCP servers.