web_search
AI agents call web_search to retrieve information from Universal AI Hub without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
Web search tools retrieve and query external data sources without modifying, deleting, or executing code. This is a Read operation. Severity is medium rather than low because: (1) an AI agent could be tricked into searching for sensitive queries, exfiltrating context or inferring private information from results, and (2) the description is empty, limiting confidence in its actual implementation details and scope.
From the tool's definition Tool name 'web_search' indicates information retrieval operation with no description provided. Based on naming convention, performs query-based data retrieval from web sources.
Attacks that exploit this kind of access
web_search. It is categorised as a Read tool in the Universal AI Hub MCP Server, which means it retrieves data without modifying state.
Register the Universal AI Hub MCP server in PolicyLayer and add a rule for web_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 Universal AI Hub. Nothing to install.
web_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 web_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 web_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.
web_search is provided by the Universal AI Hub MCP server (volkansah/universal-ai-hub). 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 →