AI agents call search-trials to retrieve information from AgentCare without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool retrieves information from ClinicalTrials.gov without modifying any data. It is a standard search/query operation with no side effects. The public nature of ClinicalTrials.gov and the read-only semantics of 'search' place this firmly in the Read category with low severity—misuse poses minimal risk as the tool only returns publicly available trial information.
From the tool's definition Tool name 'search-trials' and description 'Search ClinicalTrials.gov for relevant studies' indicate a query operation against a public clinical trials database. No modification, deletion, execution of external commands, or financial transactions are implied.
Attacks that exploit this kind of access
Search ClinicalTrials.gov for relevant studies. It is categorised as a Read tool in the AgentCare MCP Server, which means it retrieves data without modifying state.
Register the AgentCare MCP server in PolicyLayer and add a rule for search-trials: 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 AgentCare. Nothing to install.
search-trials 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 search-trials 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 search-trials. 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.
search-trials is provided by the AgentCare MCP server (kartha-ai/agentcare-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 →