AI agents use fs_write to create or update resources in Emcp — usually the action step of a workflow, after the agent has gathered context. Every call changes real data in your Emcp environment.
fs_write creates or modifies files on the filesystem. This is a Write operation (reversible modification) rather than Read (fs_read exists separately) or Destructive (fs_rm handles deletion). Severity is high because an AI agent with fs_write access could modify critical configuration files, application code, or data files, causing significant damage—though not catastrophic like Destructive operations.
From the tool's definition Tool named 'fs_write' on a filesystem operations server; sibling tools include fs_read, fs_rm, fs_mkdir, and shell_run, establishing this as a file manipulation tool.
Attacks that exploit this kind of access
fs_write. It is categorised as a Write tool in the Emcp MCP Server, which means it can create or modify data. Consider rate limits to prevent runaway writes.
Register the E MCP server in PolicyLayer and add a rule for fs_write: 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 Emcp. Nothing to install.
fs_write 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 fs_write 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 fs_write. 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.
fs_write is provided by the E MCP server (slezica/emcp). 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 →