Unclear It is not confirmed that the default lets every host through, or where a blocked result actually stops a connection.
MCP servers are add-on tools that Claude Code connects to, and each one lives at a host, the address of the machine it runs on. In cloud sessions, Claude Code now checks an MCP server's host against a list it receives from Anthropic's server. The list can be in one of three modes:
allow: only the listed hosts are permitteddeny: the listed hosts are blockedall: every host is permitted
If the list it receives is malformed, Claude Code treats it as an empty allow list, which blocks every host. A server whose host cannot be worked out is blocked too.
This applies only when CLAUDE_CODE_REMOTE is true and CLAUDE_CODE_ENVIRONMENT_KIND is unset or set to anthropic_cloud.
This gives Anthropic a way to limit which MCP servers a cloud session can reach. It has no effect when you run Claude Code on your own machine.
tengu_gleaming_creek Not enough to sayNothing here resolved what this flag was doing on this version, so nothing here should be read as on or off.
This account: no value returned · anonymous baseline: no value returned · compiled default in v2.1.305: not a boolean we can read
These values were read against a different version of Claude Code, so treat them as the nearest reading available instead of one taken on this release.
Read once, for one account on one subscription tier, against v2.1.305. It isn't a statement about your account. What a flag value here can and cannot tell you
A reading is one sample. Claude Code evaluates its flags remotely, so no client sees the targeting rule behind a value and this says nothing about your account.
A reading is one sample. Claude Code evaluates its flags remotely, so no client sees the targeting rule behind a value and this says nothing about your account.
A reading is one sample. Claude Code evaluates its flags remotely, so no client sees the targeting rule behind a value and this says nothing about your account.
A reading is one sample. Claude Code evaluates its flags remotely, so no client sees the targeting rule behind a value and this says nothing about your account.
A reading is one sample. Claude Code evaluates its flags remotely, so no client sees the targeting rule behind a value and this says nothing about your account.
It is not confirmed that the default lets every host through, or where a blocked result actually stops a connection.
New in this build: tengu_gleaming_creek