Returns output from the PTY process. Supports three modes: streaming (default) returns buffered output since last read, snapshot mode returns current terminal state, screen mode returns the parsed virtual terminal screen
AI agents call read_shell_output to retrieve information from Interactive Shell MCP without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool only reads/retrieves output from an existing shell session — it does not send input, execute commands, or modify state. However, the medium severity reflects that it operates in a persistent shell context where sensitive data (credentials, secrets, command output) could be exposed to an AI agent misusing it.
From the tool's definition 'Returns output from the PTY process' and 'returns buffered output since last read', 'returns current terminal state', 'returns the parsed virtual terminal screen'
Attacks that exploit this kind of access
Returns output from the PTY process. Supports three modes: streaming (default) returns buffered output since last read, snapshot mode returns current terminal state, screen mode returns the parsed virtual terminal screen. It is categorised as a Read tool in the Interactive Shell MCP MCP Server, which means it retrieves data without modifying state.
Register the Interactive Shell MCP server in PolicyLayer and add a rule for read_shell_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 Interactive Shell MCP. Nothing to install.
read_shell_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_shell_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_shell_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_shell_output is provided by the Interactive Shell MCP server (lightos/interactive-shell-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 →