AI
Claude Code 2.1.295 and 2.1.296 let hooks fail safe
Claude Code shipped two updates, 2.1.295 on October 8 and 2.1.296 on October 9. The headline in 2.1.295 is onFailure: "block" for command and HTTP hooks. A hook is a small script that runs when Claude Code does something. With this setting, if the hook fails to start, times out or exits unexpectedly, the action is blocked instead of going ahead. It also adds 97 fixes.
Version 2.1.296 adds CLAUDE_CODE_WORKFLOW_SUBAGENT_MODEL, which runs all workflow agents on one chosen model, plus an allow_large option for the Read tool. The MCP tool description limit rises from 2,048 to 4,096 characters. Cost tracking now reflects Sonnet 5.5 cache reads at $0.10 per million tokens.
On your desk: if any of your automation hooks guard something you care about, such as the vault or the studio's files, add onFailure: "block" so a broken hook stops the action instead of waving it through.
Read the original at Claude Code changelog →