Codex and Claude Code Changelogs

Searches for "Claude Code 3.0" are up, but no such version exists — the changelog tops out at 2.1.309, with recent real changes being Code Review levels, review stop, and ultrareview workspace trust.

A Claude Code hot update swaps the UI font to Apple's system font for a more native look, which users say resembles older Codex.

A user claims the new client uploads the git URL of the current directory, identifying org and project, with no way to disable it, linking to a test report.

Claude Code 2.1.306 ships two CLI fixes: claude -p no longer reports 'Not logged in' in cloud sessions without env vars, and MCP-heavy sessions now show original API errors instead of 'prompt is too long'.

Claude Code 2.1.302 turns on fork subagents by default across desktop, VS Code, Agent SDK and headless mode, adds a disable flag and a classifier model setting.

Claude Code 2.1.302 ships 63 CLI changes, including the Read tool refusing images over 128 MB, no prompt suggestions for credit-billed models after a credit-billed reply, and an idle compaction toggle in /config.

Tibo promised new features and optimizations for Codex within 28 days, but a week has passed with nothing notable, while Claude Code shipped 278 changes in two days, many of them exactly what Codex users want.

Claude Code released 2.1.299 and 2.1.300 back to back with 278 changes in two days, including auto-compaction, usage limit warnings in /usage, background task deduplication and sandbox escape fixes.

The Claude Code desktop update adds nested sessions that run tasks in parallel, proactive follow-up suggestions, a /btw tab for side questions, sidebar session status and recap cards, shifting the focus from writing code to managing multiple AI sessions.

Anthropic's official Code Modernization plugin guides Claude Code through legacy system migration in stages, with human confirmation at key points and per-module PROVEN/PARTLY PROVEN/NOT PROVEN verification.

Claude Code 2.1.299 adds a --disallow-tools flag, stops refusing security-testing prompts, and auto-compacts chats over 200K tokens after about 54 minutes idle, among 190 CLI changes.

REA (Reverse Engineer Anything) is an MIT-licensed MCP server and CLI that bridges coding agents like Claude Code, Cursor and Codex to tools such as Ghidra, IDA Pro and JADX, returning structured findings with evidence and confidence limits.