AI agents call read to retrieve information from Opencode without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool retrieves historical message data from remote coding sessions but does not modify, execute, or delete anything. It is purely informational with minimal blast radius if misused—an agent could only learn about existing session history, not alter system state or execute operations.
From the tool's definition The tool name is 'read' and the description states it is used to 'Read the last few messages from the most recent opencode session' and 'see what has been happening without sending a new message.' This is explicitly a retrieval operation with no side effects.
Attacks that exploit this kind of access
Read the last few messages from the most recent opencode session on an instance. Use this to see what has been happening without sending a new message. It is categorised as a Read tool in the Opencode MCP Server, which means it retrieves data without modifying state.
Register the Opencode MCP server in PolicyLayer and add a rule for read: 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 Opencode. Nothing to install.
read 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 read 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 read. 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.
read is provided by the Opencode MCP server (klutometis/opencode-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 →