MCP servers are outside programs that give Claude extra tools. Before Claude Code calls one of their tools, it now checks a list of servers sent by Anthropic through the switch tengu_gleaming_creek. If the server is on the list, Tasks is turned off for it and its tool calls run without Tasks, and a debug log line says so. Previously only a general check decided whether Tasks were used. When nothing is sent, the list is empty and no server is excluded.
MCP tool calls also gained a timeout that can be restarted while a call is running, and a second way for a call to be cancelled.
Anthropic can now switch Tasks off for a particular MCP server without affecting the others. If an MCP server's tool calls behave differently, the debug log tells you whether this list is the reason.
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.
New in this build: tengu_gleaming_creek