MCP tool idle timeout detection / Claude Code v2.1.187

What

MCP tools that stop sending progress or responses are now detected and aborted after a configurable timeout. A new env var controls the timeout duration.

Usage

# Set timeout in milliseconds (default is built-in; 0 disables timeout)
export CLAUDE_CODE_MCP_TOOL_IDLE_TIMEOUT=120000   # 2 minutes
export CLAUDE_CODE_MCP_TOOL_IDLE_TIMEOUT=0        # disable

Details

When an MCP tool produces no output for the configured duration, Claude aborts the call and reports the timeout
Error message includes the tool name and duration: "MCP server X tool Y sent no response or progress for Ns; aborting"
A new McpResponseSchemaError is thrown and surfaced when an MCP server returns a result that fails schema validation, with a description of what was wrong

Evidence

New timeout guard (search for "CLAUDE_CODE_MCP_TOOL_IDLE_TIMEOUT" or "MCP tool idle timeout")

Strings lifted out of the shipped bundle, so the claim above can be checked against them.

How sure we are

One source agreesOne thing we can check says the same as this entry.

Anthropic's release notes agreeFixed remote MCP tool calls that hang with no response for 5 minutes — they now abort with an error instead of blocking indefinitely…

See this entry in the whole of v2.1.187 →

The build it came inv2.1.187

23 Jun 2026 · 24 entries, this one is #4.

Either side of it

v2.1.186Older
v2.1.190Newer

Open the release →

Names in this entry

envCLAUDE_CODE_MCP_TOOL_IDLE_TIMEOUT
Idle timeout in milliseconds for MCP tool calls. When a stdio, HTTP, SSE, WebSocket, or claude.ai connector MCP server sends no response and no progress notification for this long, the tool call aborts with an error instead of waiting for the overall MCP_TOOL_TIMEOUT. Overrides the per-transport defaults of 300000 (5 minutes) for network servers and 1800000 (30 minutes) for stdio servers. Set to 0 to disable the idle check. Values below 1000 are raised to one second, and the value is capped at the effective MCP_TOOL_TIMEOUT. A per-server timeout in .mcp.json of at least 1000 raises that server's idle window to at least the timeout value. Doesn't apply to IDE servers or SDK in-process servers. Requires Claude Code v2.1.187 or later. Before v2.1.203, stdio servers were exempt from the idle timeout

Anthropic's note

Upstream's own release notes for v2.1.187 name something this entry names, so the two are probably about the same change. Probably: the pairing is made on a shared name rather than on the words around it.

Fixed remote MCP tool calls that hang with no response for 5 minutes — they now abort with an error instead of blocking indefinitely (override with CLAUDE_CODE_MCP_TOOL_IDLE_TIMEOUT)

Is this right?