MCP servers are add-on programs that give Claude Code extra tools, and some run long jobs called tasks that can ask you for input. When one of these input requests fails, or a server sends back a task status Claude Code cannot read, the error now lists only the fault codes. It no longer includes the error text or the input key the server supplied.
When a server asks for a kind of input Claude Code does not support, Claude Code now shows a fixed message saying so, and logs only the length of the key and the method.
An untrusted MCP server can no longer slip its own wording into Claude Code's logs or into error messages the model reads.
tengu_mcp_url_elicitation 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 v2.1.319: on
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 v2.1.319. It isn't a statement about your account. What a flag value here can and cannot tell you