Search for a regex pattern across files in the workspace. Returns matching lines with file:line prefixes.
AI agents call grep to retrieve information from Copilot Studio Code without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
The grep tool retrieves and queries data (matching lines) from files without modifying, deleting, or executing anything. It is a read-only search operation, making it a Read category risk with low severity since it only exposes existing file contents that are already accessible on the local filesystem.
From the tool's definition Tool description states: 'Search for a regex pattern across files in the workspace. Returns matching lines with file:line prefixes.' This is purely a search/query operation with no side effects.
Documented attack patterns abuse exactly the kind of access grep gives an agent:
PolicyLayer is an MCP gateway — it sits between your AI agents and Copilot Studio Code, and nothing reaches the server without passing your rules. This is the rule we recommend for grep:
{
"version": "1",
"default": "deny",
"tools": {
"grep": {}
}
}grep is read-only, so it stays allowed — but everything else on the server is denied unless you say otherwise.
Free to start. No card required.
Search for a regex pattern across files in the workspace. Returns matching lines with file:line prefixes. It is categorised as a Read tool in the Copilot Studio Code MCP Server, which means it retrieves data without modifying state.
Register the Copilot Studio Code MCP server in PolicyLayer and add a rule for grep: 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 Copilot Studio Code. Nothing to install.
grep 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 grep 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 grep. 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.
grep is provided by the Copilot Studio Code MCP server (taiki-yoshida/copilot-studio-code). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Start from Copilot Studio Code, add the rest of your stack, and see everything your agents can call. Then put policy on all of it.
Free to start. No card required.
7 Copilot Studio Code tools catalogued and risk-classified — across an index of 43,000+ MCP servers.