AI agents call read_output to retrieve information from Pentest MCP Server without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
Although the tool itself only reads data (low inherent capability), the context is critical: it retrieves output from penetration testing sessions that may contain sensitive information (credentials, system details, vulnerability data, reverse shell interactions). The blast radius is high because an AI agent could exfiltrate reconnaissance data, exploit details, or evidence of attacks.
From the tool's definition Tool name is 'read_output' and description states 'Get current output from a session' — retrieves data from a tmux session without modification.
Documented attack patterns abuse exactly the kind of access read_output gives an agent:
PolicyLayer is an MCP gateway — it sits between your AI agents and Pentest MCP Server, and nothing reaches the server without passing your rules. This is the rule we recommend for read_output:
{
"version": "1",
"default": "deny",
"tools": {
"read_output": {}
}
}read_output is read-only, so it stays allowed — but everything else on the server is denied unless you say otherwise.
Free to start. No card required.
Get current output from a session. It is categorised as a Read tool in the Pentest MCP Server MCP Server, which means it retrieves data without modifying state.
Register the Pentest MCP Server MCP server in PolicyLayer and add a rule for read_output: 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 Pentest MCP Server. Nothing to install.
read_output 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_output 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_output. 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_output is provided by the Pentest MCP Server MCP server (layesec006/pentest-mcp-server). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Start from Pentest MCP Server, 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.
12 Pentest MCP Server tools catalogued and risk-classified — across an index of 43,000+ MCP servers.