Read a task file by type (init, plan, done, error)
AI agents call read_task to retrieve information from Agent Communication MCP Server without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool retrieves task information from existing task files based on their type/state. It queries and returns data without side effects, making it a straightforward Read operation. The low severity reflects minimal blast radius: reading task metadata poses negligible risk even if called indiscriminately by an agent.
From the tool's definition Tool name is 'read_task' and description states 'Read a task file by type (init, plan, done, error)' — explicit read operation with no modification or deletion.
Attacks that exploit this kind of access
Read a task file by type (init, plan, done, error). It is categorised as a Read tool in the Agent Communication MCP Server MCP Server, which means it retrieves data without modifying state.
Register the Agent Communication MCP Server MCP server in PolicyLayer and add a rule for read_task: 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 Agent Communication MCP Server. Nothing to install.
read_task 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_task 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_task. 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_task is provided by the Agent Communication MCP Server MCP server (jerfowler/agent-comm-mcp-server). 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 →