AI agents call market.arb.correlation to retrieve information from Rekko MCP without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
The tool appears designed to retrieve or analyze correlation data for arbitrage detection—a read-level intelligence operation. However, it operates in a financial prediction market context where misuse could inform high-value trading decisions. While technically a Read operation, the high blast radius (informing large financial decisions) elevates severity to high.
From the tool's definition Tool is part of market.arb.* family on a prediction market platform; the 'arb' prefix indicates arbitrage detection, which is a research and analysis function. Sibling tools like market.arb.get and market.arb.live are market intelligence functions.
Attacks that exploit this kind of access
market.arb.correlation. It is categorised as a Read tool in the Rekko MCP MCP Server, which means it retrieves data without modifying state.
Register the Rekko MCP server in PolicyLayer and add a rule for market.arb.correlation: 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 Rekko MCP. Nothing to install.
market.arb.correlation is a Read tool with low risk. Read-only tools are generally safe to allow by default.
Yes. Add a rate_limit block to the market.arb.correlation 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 market.arb.correlation. 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.
market.arb.correlation is provided by the Rekko MCP server (rekko-ai/rekko-mcp). 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 →