Claude Code v2.1.295: Hooks Can Now Block on Failure with onFailure: "block"

Hi, it's Shii-chan! A new Claude Code release, v2.1.295, is here, and I'm picking out the highlights!

What was announced?

The Claude Code Release page on GitHub published v2.1.295. It's a big one, with added features, improvements, changes, and a very long list of fixes.

Until now, a hook that couldn't start, timed out, or exited with an unexpected code could let the action go through anyway.

If you use hooks as guards, an action slipping through when the guard didn't run is a worry.

Here are the notable additions!

onFailure: "block": for command and HTTP hooks, a hook that can't start, times out, or exits with an unexpected code blocks the action instead of letting it through
Program Status Protocol (OSC 7501) support: terminals that implement it can show whether Claude Code is working, waiting on you, or done
Quoted text in the /copy picker: a drafted message copies without its > markers
A warning from claude plugin install, enable, disable, and marketplace add when the settings file they write to does not load
A line on stderr (when it is a terminal) saying what a claude -p run is waiting for when it stays open after its last turn
CLAUDE_CODE_RETRY_WATCHDOG_MAX_WAIT_MS: limits how long unattended retry mode waits out 429 and 529 errors

There are many additions around the Claude apps gateway too:

timeouts.upstream_ttfb_ms support for Bedrock, Vertex, Foundry, and other cloud upstreams: a value you set limits how long a stream may take to start, after which it fails over or gets a 502
An optional models list on every upstream: only the listed models are sent there (on failover too), and one * in an entry is a wildcard
forceLoginMethod: "gateway" and forceLoginGatewayUrl in your own user settings on machines with no managed settings
upstream_request_id in the inference audit event, for support cases
A request-id header on successful inference responses, so request_id in Claude Code telemetry matches the gateway's audit log
Token counts behind a gateway on Amazon Bedrock now come from AWS's CountTokens API. Grant bedrock:CountTokens to use it

For mods, $.ui.notify and children for Button (strings and Text) were added.

Some behavior changes:

MCP tool descriptions loaded through tool search are now cut at 16,384 characters instead of 2,048
Subagents preload at most 32 skills from the skills field
WebSocket (ws) MCP servers: a message over 16 MiB is no longer parsed and closes the connection, matching the limit of the other transports

The fixes cover a lot of ground. A few examples:

Every request failing on a [1m] model when a gateway, Bedrock, Vertex, or Foundry refuses the context-1m beta. Claude Code now resends without it
Remote MCP servers staying disconnected after an outage, or reconnecting in a tight loop. Repeated drops now back off, up to 30s
The terminal freezing when a response ran to tens of thousands of lines
Project-scope plugins not loading on Windows due to drive-letter casing

VS Code also gets a chat row for files Claude sends you; click the file name to open it in the editor.

This release is for anyone building guards with hooks or running the Claude apps gateway!