AI agents call nrf_search to retrieve information from Nrf without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool performs information retrieval across a code repository without creating, modifying, deleting, or executing any operations. It is purely a search and query function that returns results without side effects. The low severity reflects that misuse would only expose existing documentation or code that is already in the repository, with no capability to alter, destroy, or execute arbitrary operations.
From the tool's definition Tool description states it 'Search for code or documentation across the nRF Connect SDK repo' and 'Supports GitHub search qualifiers to narrow results'.
Attacks that exploit this kind of access
Search for code or documentation across the nRF Connect SDK repo using GitHub code search. Supports GitHub search qualifiers to narrow results: - Plain keyword:. It is categorised as a Read tool in the Nrf MCP Server, which means it retrieves data without modifying state.
Register the Nrf MCP server in PolicyLayer and add a rule for nrf_search: 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 Nrf. Nothing to install.
nrf_search 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 nrf_search 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 nrf_search. 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.
nrf_search is provided by the Nrf MCP server (pshanesmith/nrf-mcp). 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 →