Change to directory, list its contents, get the notes from CLAUDE.md.
AI agents call work_on to retrieve information from MCP File System Server without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool combines directory navigation and content inspection without modifying, executing, or deleting anything. The most severe action is reading a file, which is a Read category operation. Confidence is high because the description is explicit and the functionality is clearly informational.
From the tool's definition Tool performs three read-only operations: 'Change to directory' (navigation state), 'list its contents' (directory enumeration via ls), and 'get the notes from CLAUDE.md' (file read). No write, execute, or destructive operations are described.
Attacks that exploit this kind of access
Change to directory, list its contents, get the notes from CLAUDE.md. It is categorised as a Read tool in the MCP File System Server MCP Server, which means it retrieves data without modifying state.
Register the MCP File System Server MCP server in PolicyLayer and add a rule for work_on: 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 MCP File System Server. Nothing to install.
work_on 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 work_on 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 work_on. 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.
work_on is provided by the MCP File System Server MCP server (kvas-it/mcp-server-fs). 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 →