io.github.ClawCall-Dev/clawcall

It requires CLAWCALL_API_KEY and CLAWCALL_BASE_URL, both sensitive configurations that enable access to a remote service; if exposed, they could allow service abuse or redirect requests to unintended destinations. The materials do not describe key scope, least-privilege controls, or storage handling.

Its stated purpose is AI-powered US phone calling, and the required BASE_URL indicates it will make outbound requests to a remote service and may transmit call-related content. No fixed remote host is listed and the README is absent, so the actual destinations, transmitted data types, and whether a custom BASE_URL can alter egress are unclear.

The system flags it as executes-code, meaning the MCP can start local processes or execute code on the host; this is a normal but privileged capability for such tools and should be sandboxed. The available materials do not show requests for abnormal system permissions beyond its calling use case, so this alone does not justify a high-risk rating.

The materials do not specify which local files, logs, contacts, or call records it can read or write, nor do they define a least-privilege data boundary. Because it can execute code and is used for phone interactions, it should be assumed to potentially touch user inputs and runtime data, but there is no concrete evidence of clearly excessive access.

It comes from the official Registry, has a public source repository, and was updated within the last year, which are positive signs for auditability and ongoing maintenance. Although community adoption is low (0 stars) and the license is unspecified, that reduces maturity confidence rather than constituting a clear supply-chain red flag.