Claude Code 2.1.294 fixes instruction-style hooks and premature agent stops
ClaudeCodeLog · x · 2026-10-09
Claude Code shipped v2.1.294, roughly 20 hours after 2.1.293. The key fix: prompt/agent hooks written as natural-language instructions (e.g. "Block commands that...") were previously treated as regular instructions, letting through commands they were meant to block. It also improves how instruction-style hooks on Stop and SubagentStop (e.g. "Carry on if the build is broken") are judged, making Claude less likely to stop early. Prompt telemetry: +3 prompt files and +8,477 tokens (+16.1%) vs 2.1.293, with system prompts at 80% of the mix.
More from coding & agent
- Developer Now Does Half His Coding From His Phone Using AI Workspaces — lucasmeijer · 2026-10-09
- Dev shares agent workflow: background workspaces protect your focus — lucasmeijer · 2026-10-09
- Dev shows remote workspaces running any agent on any model, incl. vanilla Claude Code — lucasmeijer · 2026-10-09
- Open-source lithos-metal hits 200+ tokens/s/user on Qwen3.8-27B with one M5 Max — JiaZhihao · 2026-10-09
- Claude Code 2.1.294 fixes instruction-based hooks that failed to block commands and premature agent stops — ClaudeCodeLog · 2026-10-09
- CodeRabbit hits 1,049,817 open-source code reviews in September, 2.49x April's volume — leslysandra · 2026-10-09