Release 0.123.0-cometix · Haleclipse/codex

release: rust-v0.123.0-cometix

New Features

Added a built-in amazon-bedrock model provider with configurable AWS profile support (openai#18744).
Added /mcp verbose for full MCP server diagnostics, resources, and resource templates while keeping plain /mcp fast (openai#18610).
Made plugin MCP loading accept both mcpServers and top-level server maps in .mcp.json (openai#18780).
Improved realtime handoffs so background agents receive transcript deltas and can explicitly stay silent when appropriate (openai#18597, openai#18761, openai#18635).
Added host-specific remote_sandbox_config requirements for remote environments (openai#18763).
Refreshed bundled model metadata, including the current gpt-5.4 default (openai#18586, openai#18388, openai#18719).

Fixed /copy after rollback so it copies the latest visible assistant response, not a pre-rollback response (openai#18739).
Queued normal follow-up text submitted while a manual shell command is running, preventing stuck Working states (openai#18820).
Fixed Unicode/dead-key input in VS Code WSL terminals by disabling the enhanced keyboard mode there (openai#18741).
Prevented stale proxy environment variables from being restored from shell snapshots (openai#17271).
Made codex exec inherit root-level shared flags such as sandbox and model options (openai#18630).
Removed leaked review prompts from TUI transcripts (openai#18659).

Added and tightened the Code Review skill instructions used by Codex-driven reviews (openai#18746, openai#18818).
Documented intentional await-across-lock cases and enabled Clippy linting for them (openai#18423, openai#18698).
Updated app-server protocol docs for threadless MCP resource reads and namespaced dynamic tools (openai#18292, openai#18413).

Fixed high-severity dependency alerts by pinning patched JS and Rust dependencies (openai#18167).
Reduced Rust dev build debug-info overhead while preserving useful backtraces (openai#18844).
Refreshed generated Python app-server SDK types from the current schema (openai#18862).

Session module restructured; token_info_and_rate_limits migrated to new session path
Upstream now sends token usage after resume (merged with cometix rate limit notification)
Restored 8 legacy models removed by upstream
Added free plan visibility to gpt-5.4 and gpt-5.3-codex
Standalone installer variants redirect to npm install for cometix fork
Removed upstream-only workflows (dotslash, v8-canary, zsh, argument-comment-lint, etc.)