AI agents invoke run_ab_test to trigger actions in Guru Pk. 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/B testing operations that compare different modes. While the tests themselves are not destructive (results are temporary, insights only), the tool triggers computational processes with side effects (test execution, data collection, metric calculation). It is not a simple read (which would only retrieve pre-existing test results), but rather executes test logic based on arguments.
From the tool's definition Tool name is 'run_ab_test' and description states it '运行A/B测试' (runs A/B tests), which involves executing comparative testing operations between sequence mode and batch processing mode.
Documented attack patterns abuse exactly the kind of access run_ab_test gives an agent:
PolicyLayer is an MCP gateway — it sits between your AI agents and Guru Pk, and nothing reaches the server without passing your rules. This is the rule we recommend for run_ab_test:
{
"version": "1",
"default": "deny",
"tools": {
"run_ab_test": {
"limits": [
{
"counter": "run_ab_test_rate",
"window": "minute",
"max": 10,
"scope": "grant"
}
]
}
}
}run_ab_test 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.
运行A/B测试对比序列模式和批处理模式. It is categorised as a Execute tool in the Guru Pk MCP Server, which means it can trigger actions or run processes. Use rate limits and argument validation.
Register the Guru Pk MCP server in PolicyLayer and add a rule for run_ab_test: 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 Guru Pk. Nothing to install.
run_ab_test 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 run_ab_test 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 run_ab_test. 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.
run_ab_test is provided by the Guru Pk MCP server (mitsudoai/guru-pk-mcp). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Start from Guru Pk, 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.
22 Guru Pk tools catalogued and risk-classified — across an index of 43,000+ MCP servers.