AI agents invoke start_game to trigger actions in Gomoku. 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.
The tool executes a game state transition command whose effects depend on room context (presence of 2 players). While the blast radius is low (isolated to a game room with no persistent damage), it is not Read (no query), not Write (irreversible state change that initiates gameplay), and not Destructive/Financial (no data loss or money involved).
From the tool's definition Tool name 'start_game' combined with description '开始游戏(需要房间内有2名玩家)' (Start game [requires 2 players in room]) indicates it triggers game initiation logic.
Attacks that exploit this kind of access
开始游戏(需要房间内有2名玩家). It is categorised as a Execute tool in the Gomoku MCP Server, which means it can trigger actions or run processes. Use rate limits and argument validation.
Register the Gomoku MCP server in PolicyLayer and add a rule for start_game: 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 Gomoku. Nothing to install.
start_game 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_game 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_game. 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_game is provided by the Gomoku MCP server (sanshao85/gomoku-ai-battle). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Every MCP server has a record like this.
Type a name, get the same breakdown: verified identity, auth posture, risk grade, capabilities, recommended policy.
Teams ship this data inside their own products. See what a licence covers →