transcribe_once
AI agents call transcribe_once to retrieve information from Claude Voice Commands without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool retrieves/converts audio input into text output with no side effects on data or systems. While it involves processing audio, the operation is fundamentally a data retrieval/transformation task (speech-to-text conversion).
From the tool's definition Tool name 'transcribe_once' combined with server description indicating 'speech-to-text' functionality suggests audio capture and conversion to text. The name implies a single, one-time transcription operation without modification or deletion of data.
Attacks that exploit this kind of access
transcribe_once. It is categorised as a Read tool in the Claude Voice Commands MCP Server, which means it retrieves data without modifying state.
Register the Claude Voice Commands MCP server in PolicyLayer and add a rule for transcribe_once: 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 Claude Voice Commands. Nothing to install.
transcribe_once 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 transcribe_once 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 transcribe_once. 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.
transcribe_once is provided by the Claude Voice Commands MCP server (jwitcoff/claude_chat). 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 →