MCP tools now pass up to 500K via _meta annotation; plugins gain bin/ executables; --resume transcript breaks and plan-mode container restart failures fixed.
_meta["anthropic/maxResultSizeChars"] annotation (up to 500K), allowing larger results like DB schemas to pass through without truncationdisableSkillShellExecution setting to disable inline shell execution in skills, custom slash commands, and plugin commands/claude-api skill guidance for agent design patterns including tool surface decisions, context management, and caching strategystripAnsi on Bun by routing through Bun.stripANSI--resume that could lose conversation history when async transcript writes fail silentlycmd+delete not deleting to start of line on iTerm2, kitty, WezTerm, Ghostty, and Windows Terminalpermissions.defaultMode: "auto" in settings.jsonclaude-cli://open?q= deep links (encoded newlines %0A no longer rejected)bin/ and invoke them as bare commands from the Bash tool/feedback now explains why it's unavailable instead of disappearing from the slash menuold_string anchors, reducing output tokens