Retell AI

It requires RETELL_API_KEY, a sensitive credential that can authorize access to a third-party voice/agent service; if exposed, it could enable account misuse or unwanted charges. The materials do not indicate requests for credentials beyond its stated function.

As a tool for voice agent phone calls, agent management, LLMs, and voice discovery, it would normally need to communicate with Retell-related services; the materials do not list specific remote hosts, so egress transparency is limited, but there is no evidence of unrelated or suspicious endpoints.

The system checks explicitly mark executes-code, meaning it runs locally as an MCP server process. This is a standard capability for this class of tool; the materials do not show requests for elevated system privileges, persistence, or actions clearly unrelated to its stated purpose.

The materials do not specify what local files, databases, or other resources it can read or write, so data access boundaries are not fully clear. Based on its voice-agent and management functions, it may process user inputs and service-related data, but there is no indication of broad local file access or obviously excessive permissions.

The source is an official registry entry with an auditable open-source repository and updates within the last year; these are positive signals that reduce risk. Although the license is not stated and community stars are low, there are no signs of a closed-source, abandoned, or otherwise clearly risky supply-chain profile.