inspect_untrusted_text

What inspect_untrusted_text does on Project Guardian MCP

AI agents call inspect_untrusted_text to retrieve information from Project Guardian MCP without modifying anything. It is typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.

Why inspect_untrusted_text is rated Low

This tool analyzes input text to identify security patterns but does not modify data, execute code, delete information, or trigger external operations. It is a read-only security analysis capability that examines content for malicious patterns. The low severity reflects that misuse would primarily affect analysis accuracy rather than system state or external consequences.

From the tool's definition Tool performs text normalization and detection of prompt-injection indicators; the verb 'inspect' and 'detect' indicate passive analysis without data modification, deletion, or external execution. No side effects mentioned.

The rule that runs inspect_untrusted_text safely

PolicyLayer is an MCP gateway: it sits between your AI agents and Project Guardian MCP, and checks every tool call against a rule you set before the call runs. Nothing changes on the server itself. For inspect_untrusted_text, this is the rule to start with:

{
  "version": "1",
  "default": "deny",
  "tools": {
    "inspect_untrusted_text": {}
  }
}

Questions about inspect_untrusted_text