AI agents use download_audio to create or update resources in Claud-Ear — usually the action step of a workflow, after the agent has gathered context. Every call changes real data in your Claud-Ear environment.
This tool fetches remote audio content and writes it to the local filesystem as a WAV file. It creates new data (the downloaded file) on disk, making it a Write operation. It is not purely Read (it has the side effect of writing a file), not Execute (no code/commands are run), and not Destructive. Misuse could involve downloading large files, copyrighted content, or filling disk space, giving it medium severity.
From the tool's definition Download audio from a YouTube or SoundCloud URL as a WAV file
Attacks that exploit this kind of access
Download audio from a YouTube or SoundCloud URL as a WAV file. It is categorised as a Write tool in the Claud-Ear MCP Server, which means it can create or modify data. Consider rate limits to prevent runaway writes.
Register the Claud-Ear MCP server in PolicyLayer and add a rule for download_audio: 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 Claud-Ear. Nothing to install.
download_audio is a Write tool with medium risk. Write tools should be rate-limited to prevent accidental bulk modifications.
Yes. Add a rate_limit block to the download_audio 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 download_audio. 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.
download_audio is provided by the Claud-Ear MCP server (null-phnix/claud-ear). 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 →