Release 0.145.0-cometix · Haleclipse/codex

New Features

Added experimental paginated thread history with efficient resume, search, persisted names, sub-agent support, and memories. (openai#33364, openai#33907, openai#34085, openai#34229, openai#34386)
Expanded /import to migrate Cursor and Claude Code settings, MCP servers, plugins, sessions, commands, and project-scoped memories. (openai#31672, openai#33411, openai#33426, openai#33444)
Added experimental Amazon Bedrock login, custom endpoint and authentication support, and GPT-5.6 Sol as the default Bedrock model. (openai#31327, openai#33170, openai#33175, openai#32288, openai#33695)
Added audio inputs and tool outputs, including common local audio formats, and introduced streaming realtime V3 conversations. (openai#33261, openai#33856, openai#33932, openai#34080, openai#34385)
Stabilized the opt-in multi-agent V2 experience with configurable sub-agent models, reasoning levels, concurrency, restored roles, and improved agent navigation. (openai#33550, openai#33631, openai#33657, openai#33841, openai#34383)
Added secure, clickable inline visualization links in the terminal UI. (openai#33925, openai#34217, openai#34346)

Editing an earlier prompt or retrying a safety-buffered turn now creates a contextual branch, preserving the original conversation, attachments, and mention bindings. (openai#33201, openai#33207, openai#33211)
Improved terminal responsiveness for long conversations and streamed output through incremental Markdown rendering, fewer redraws, caching, and bounded command output. (openai#34045, openai#34049, openai#34216, openai#34223, openai#34359)
Prevented slow or conflicting MCP startup and authentication flows by enforcing startup timeouts, avoiding blocking OAuth discovery, serializing refreshes, and reusing tool catalogs safely. (openai#32229, openai#32781, openai#32825, openai#33184, openai#33297)
Improved Windows execution and sandbox reliability, including native exec-server sandboxing, network-proxy enforcement, hidden helper consoles, and correctly quoted hook commands. (openai#32849, openai#32857, openai#33926, openai#34423)
Fixed compact release-metadata parsing and macOS code-mode installation, with an in-process fallback when the external code-mode host is unavailable. (openai#31667, openai#31876, openai#31899)
Strengthened safety and approval handling with better forced-rm detection, consistent full-access confirmation, and preserved rejection reasons across tools. (openai#32989, openai#33464, openai#34400)

Updated the bundled OpenAI Docs skill with current GPT-5.6 model resolution, prompting, and migration guidance across macOS, Linux, and Windows. (openai#31842, openai#33121)

Migrated bundled GPT-5.4 selections and internal uses to the corresponding GPT-5.6 Terra and Luna variants. (openai#33173)
Reduced startup and large-context overhead with concurrent skill/plugin discovery and more efficient remote compaction. (openai#31566, openai#33369, openai#33423, openai#34431)
Updated the packaged ripgrep binary to 15.2.0. (openai#34384)