v0.4.2

Added

• max reasoning effort level for both codex and codex-reply tools
• Known models hint in model parameter description (gpt-5.6-sol) so LLM agents can resolve "latest model" requests correctly

🤖 Generated with Claude Code