{"slug":"frontier-ai-labs","name":"Frontier AI Labs","description":"Frontier AI research labs and model providers, side by side.","orgs":[{"slug":"anthropic","name":"Anthropic","domain":"anthropic.com","avatarUrl":null,"description":null,"githubHandle":"anthropics"},{"slug":"openai","name":"OpenAI","domain":"openai.com","avatarUrl":null,"description":null,"githubHandle":"openai"}],"releases":[{"id":"rel_oqWmfXgDrxV_twO3Kygss","version":"v2.1.137","type":"feature","title":"v2.1.137","summary":"## What's changed\n\n- [VSCode] Fixed extension failing to activate on Windows\n","content":"## What's changed\n\n- [VSCode] Fixed extension failing to activate on Windows\n","publishedAt":"2026-05-09T00:11:04.000Z","url":"https://code.claude.com/docs/en/changelog#2-1-137","media":[],"prerelease":false,"source":{"slug":"claude-code","name":"Claude Code","type":"scrape"},"org":{"slug":"anthropic","name":"Anthropic"},"product":{"slug":"claude-code","name":"Claude Code"}},{"id":"rel_DVdEV7-p7JjPVOdCPfJTc","version":"2.1.136","type":"feature","title":"2.1.136","summary":"- Added `CLAUDE_CODE_ENABLE_FEEDBACK_SURVEY_FOR_OTEL` to re-enable the session quality survey for enterprises capturing responses through OpenTelemetr...","content":"- Added `CLAUDE_CODE_ENABLE_FEEDBACK_SURVEY_FOR_OTEL` to re-enable the session quality survey for enterprises capturing responses through OpenTelemetry\n\n- Added `settings.autoMode.hard_deny` for auto mode classifier rules that block unconditionally regardless of user intent or allow exceptions\n\n- Fixed MCP servers configured in `.mcp.json`, plugins, and claude.ai connectors silently disappearing after `/clear` in the VS Code extension, JetBrains plugin, and Agent SDK\n\n- Fixed a rare login loop where a concurrent credential write could overwrite a freshly-rotated OAuth token and force re-login\n\n- Fixed MCP OAuth refresh tokens being lost when multiple servers refresh concurrently — users with several remote MCP servers should no longer need daily re-authentication\n\n- Fixed an API error (400) when extended thinking emitted a redacted thinking block after a tool call\n\n- Fixed `--resume` / `--continue` not finding sessions when the project path contains underscores\n\n- Fixed plan mode not blocking file writes when a matching `Edit(...)` allow rule exists\n\n- WSL2: image paste from Windows clipboard now works via a PowerShell fallback when xclip/wl-paste cannot read image data\n\n- Fixed plugin `Stop`/`UserPromptSubmit` hooks failing when cache cleanup deletes a version still in use by a running session\n\n- Improved visual consistency across slash command dialogs: standardized footer hints, dialog spacing, and arrow-key styling, and the dialog frame now appears immediately during loading instead of popping in after\n\n- Fixed colors appearing at wrong positions in bash command output and markdown code blocks\n\n- Fixed ReasonML diffs rendering corrupted \"undefined\" text artifacts at word-diff boundaries\n\n- Fixed worktree exit dialog warning about uncommitted files in the wrong directory after worktree removal\n\n- Fixed `@` file picker not matching files created mid-session in small non-git directories\n\n- Fixed `@`-mention file picker not finding files in directories with more than 100 entries\n\n- Fixed failed tool calls not being click-to-expand in fullscreen mode when their output was truncated\n\n- Fixed Backspace and Ctrl+Backspace getting swapped after using Ctrl+G to open an external editor on terminals with persistent extended-key modes\n\n- Fixed `/usage` weekly reset showing time of day instead of the calendar date\n\n- Fixed welcome banner ellipsis causing column overflow on CJK terminals\n\n- Fixed `/insights` crash when session history contains tool calls with malformed input fields\n\n- Fixed a renderer crash when a tool's collapsibility classification changes mid-session\n\n- Fixed a `skills` entry in `plugin.json` hiding the plugin's default `skills/` directory, and listing a file path now shows an error instead of failing silently\n\n- Fixed IDE shell-integration lock files not respecting `CLAUDE_CONFIG_DIR`\n\n- Fixed trailing whitespace in copied terminal output during streaming\n\n- Fixed plugin uninstall and enable/disable not matching slugs case-insensitively\n\n- Fixed tool error truncation marker showing a negative count for surrogate-pair strings\n\n- Fixed env vars from `CLAUDE_ENV_FILE` SessionStart hooks going stale after `/resume` or `/clear`\n\n- Fixed `/branch` saving a multi-line session title when given a pasted multi-line name\n\n- Fixed a stray leading space on the second line of wrapped text at the column boundary\n\n- Fixed Esc not dismissing dialogs in `/install-github-app`, `/desktop`, `/resume`, and `/web-setup`\n\n- Fixed `/doctor` MCP schema errors not naming the missing field or showing the source file path\n\n- Fixed Bash permission prompts showing an internal parser diagnostic instead of a user-readable explanation\n\n- Fixed plugin slash commands with spaces (e.g. `/myplugin review`) not resolving to their namespaced form\n\n- Fixed `AskUserQuestion` discarding multi-select answers when supplied as an array\n\n- Fixed `/clear ` not labeling the cleared session for `/resume`\n\n- Fixed `CronList` output missing qualifiers and the scheduled prompt\n\n- Fixed \"Jump to bottom\" overlay leaving color artifacts on CJK characters in fullscreen mode\n\n- Fixed wide markdown tables leaving a stale bordered render in terminal scrollback while streaming\n\n- Fixed pasted text being silently dropped when a long prompt with a pasted-text placeholder was auto-truncated\n\n- Fixed `/release-notes` getting stuck on an old version after a failed changelog refresh\n\n- Fixed `/mcp` server list not scrolling when there are more servers than fit in the terminal\n\n- Fixed mid-input slash command autocomplete not working after an initial slash command\n\n- Fixed scrolling to bottom re-engaging auto-follow with `autoScrollEnabled: false`\n\n- Fixed prompt suggestions being auto-submitted by Enter on an empty input instead of requiring Tab or arrow to accept\n\n- Fixed keyboard shortcut hints not reflecting rebound keys from `keybindings.json`\n\n- Fixed `/settings` language change being reverted on Escape after confirming\n\n- Fixed `/terminal-setup` only appearing in autocomplete on exact name match instead of partial prefixes\n\n- Fixed \"Chat about this\" on an `AskUserQuestion` dialog erasing the question text\n\n- Fixed MCP tool results being invisible when the server returns content blocks\n\n- Improved error message when `--worktree` collides with an existing or stale worktree\n\n- Changed plugin marketplace removal key to `d` (matching delete elsewhere) instead of `r` which collided with retry","publishedAt":"2026-05-08T18:47:30.000Z","url":"https://code.claude.com/docs/en/changelog#2-1-136","media":[],"prerelease":false,"source":{"slug":"claude-code","name":"Claude Code","type":"scrape"},"org":{"slug":"anthropic","name":"Anthropic"},"product":{"slug":"claude-code","name":"Claude Code"}},{"id":"rel_G0qD0bveCAlDzXjDY2sK6","version":"0.130.0","type":"feature","title":"Codex CLI Release: 0.130.0","summary":"## New Features\n\n- Plugin details now show bundled hooks, and plugin sharing exposes link metadata plus discoverability controls. ([#21447](https://gi...","content":"## New Features\n\n- Plugin details now show bundled hooks, and plugin sharing exposes link metadata plus discoverability controls. ([#21447](https://github.com/openai/codex/pull/21447), [#21495](https://github.com/openai/codex/pull/21495), [#21637](https://github.com/openai/codex/pull/21637))\n\n- Added `codex remote-control` as a simpler entrypoint for starting a headless, remotely controllable app-server. ([#21424](https://github.com/openai/codex/pull/21424))\n\n- App-server clients can page large threads with unloaded, summary, or full turn item views. ([#21566](https://github.com/openai/codex/pull/21566))\n\n- Bedrock auth can now use AWS console-login credentials from `aws login` profiles. ([#21623](https://github.com/openai/codex/pull/21623))\n\n- `view_image` can resolve files through the selected environment for multi-environment sessions. ([#21143](https://github.com/openai/codex/pull/21143))\n\n## Bug Fixes\n\n- Live app-server threads now pick up config changes without requiring a restart. ([#21187](https://github.com/openai/codex/pull/21187))\n\n- Turn diffs stay accurate across apply-patch operations, including partial failures that still mutated files. ([#21180](https://github.com/openai/codex/pull/21180), [#21518](https://github.com/openai/codex/pull/21518))\n\n- Thread summaries, renames, resume, and fork paths work better through `ThreadStore`, including threads without local rollout paths. ([#21264](https://github.com/openai/codex/pull/21264), [#21265](https://github.com/openai/codex/pull/21265), [#21266](https://github.com/openai/codex/pull/21266))\n\n- Remote compaction now emits `response.processed` for v2 streams and avoids sending `service_tier` on API-key compact requests. ([#21642](https://github.com/openai/codex/pull/21642), [#21676](https://github.com/openai/codex/pull/21676))\n\n- Windows sandbox setup now grants sandbox users access to the desktop runtime binary cache. ([#21564](https://github.com/openai/codex/pull/21564))\n\n- Removed stale “research preview” wording from the `codex exec` startup banner. ([#21683](https://github.com/openai/codex/pull/21683))\n\n## Documentation\n\n- Fixed issue templates so CLI reports keep the intended guidance, labels apply correctly, and feature requests link to the right contributing docs. ([#21685](https://github.com/openai/codex/pull/21685), [#21686](https://github.com/openai/codex/pull/21686), [#21688](https://github.com/openai/codex/pull/21688))\n\n- Updated install and tooling docs to consistently use `cargo install --locked`. ([#21592](https://github.com/openai/codex/pull/21592))\n\n## Chores\n\n- Added a faster Cargo profiling build profile and disabled empty doctest targets to speed up Rust development loops. ([#21574](https://github.com/openai/codex/pull/21574), [#21584](https://github.com/openai/codex/pull/21584))\n\n- Hardened dependency and CI hygiene with fully qualified GitHub Action pins, a Dependabot cooldown, and a `cargo-shear` upgrade. ([#21436](https://github.com/openai/codex/pull/21436), [#21547](https://github.com/openai/codex/pull/21547), [#21599](https://github.com/openai/codex/pull/21599))\n\n- Simplified internal surfaces by removing unused device-key APIs, extra skills roots, the remote thread-store implementation, and string-keyed MCP tool maps. ([#21487](https://github.com/openai/codex/pull/21487), [#21485](https://github.com/openai/codex/pull/21485), [#21596](https://github.com/openai/codex/pull/21596), [#21454](https://github.com/openai/codex/pull/21454))\n\n- Added configurable OpenTelemetry trace metadata and richer review/feedback analytics for better debugging and triage. ([#21556](https://github.com/openai/codex/pull/21556), [#18747](https://github.com/openai/codex/pull/18747), [#21434](https://github.com/openai/codex/pull/21434), [#21498](https://github.com/openai/codex/pull/21498))\n\n## Changelog\n\nFull Changelog: [rust-v0.129.0...rust-v0.130.0](https://github.com/openai/codex/compare/rust-v0.129.0...rust-v0.130.0)\n\n- [#21494](https://github.com/openai/codex/pull/21494) [codex] fix PluginListParams test initializer [@xli-oai](https://github.com/xli-oai)\n\n- [#21447](https://github.com/openai/codex/pull/21447) Show plugin hooks in plugin details [@abhinav-oai](https://github.com/abhinav-oai)\n\n- [#21356](https://github.com/openai/codex/pull/21356) feat: make built-in MCPs first-class runtime servers [@jif-oai](https://github.com/jif-oai)\n\n- [#21180](https://github.com/openai/codex/pull/21180) Make turn diff tracking operation backed [@jif-oai](https://github.com/jif-oai)\n\n- [#21498](https://github.com/openai/codex/pull/21498) [codex] add account id to feedback uploads [@pakrym-oai](https://github.com/pakrym-oai)\n\n- [#21487](https://github.com/openai/codex/pull/21487) device-key: clean up unused crate [@euroelessar](https://github.com/euroelessar)\n\n- [#21518](https://github.com/openai/codex/pull/21518) fix: preserve exact turn diffs after partial apply_patch failures [@jif-oai](https://github.com/jif-oai)\n\n- [#18747](https://github.com/openai/codex/pull/18747) [codex-analytics] add tool review event schema [@rhan-oai](https://github.com/rhan-oai)\n\n- [#21495](https://github.com/openai/codex/pull/21495) feat: Expose plugin share metadata in shareContext [@xl-openai](https://github.com/xl-openai)\n\n- [#21454](https://github.com/openai/codex/pull/21454) [codex] Remove string-keyed MCP tool maps [@pakrym-oai](https://github.com/pakrym-oai)\n\n- [#21424](https://github.com/openai/codex/pull/21424) add top-level remote-control command [@owenlin0](https://github.com/owenlin0)\n\n- [#21187](https://github.com/openai/codex/pull/21187) app-server: refresh live threads from latest config snapshot [@jif-oai](https://github.com/jif-oai)\n\n- [#21461](https://github.com/openai/codex/pull/21461) [codex] Move tool specs onto handlers [@pakrym-oai](https://github.com/pakrym-oai)\n\n- [#21547](https://github.com/openai/codex/pull/21547) Upgrade `cargo-shear` to 1.11.2 [@charliemarsh-oai](https://github.com/charliemarsh-oai)\n\n- [#21264](https://github.com/openai/codex/pull/21264) Move thread name edits to ThreadStore [@wiltzius-openai](https://github.com/wiltzius-openai)\n\n- [#21266](https://github.com/openai/codex/pull/21266) [codex] Fix pathless thread summaries [@wiltzius-openai](https://github.com/wiltzius-openai)\n\n- [#21265](https://github.com/openai/codex/pull/21265) Route ThreadManager rollout path reads through thread store [@wiltzius-openai](https://github.com/wiltzius-openai)\n\n- [#21564](https://github.com/openai/codex/pull/21564) Grant sandbox users access to desktop runtime bin [@iceweasel-oai](https://github.com/iceweasel-oai)\n\n- [#21582](https://github.com/openai/codex/pull/21582) Use descriptive names for Cargo profile options [@zanie-oai](https://github.com/zanie-oai)\n\n- [#21574](https://github.com/openai/codex/pull/21574) Add a Cargo build profile for benchmarking [@zanie-oai](https://github.com/zanie-oai)\n\n- [#21436](https://github.com/openai/codex/pull/21436) [codex] Fully qualify hash-pins in GitHub Actions [@ww-oai](https://github.com/ww-oai)\n\n- [#21592](https://github.com/openai/codex/pull/21592) Ensure all mentions of cargo-install are --locked [@gankra-oai](https://github.com/gankra-oai)\n\n- [#21584](https://github.com/openai/codex/pull/21584) Disable empty Cargo test targets [@charliemarsh-oai](https://github.com/charliemarsh-oai)\n\n- [#21566](https://github.com/openai/codex/pull/21566) feat(app-server, threadstore): Thread pagination APIs and ThreadStore contract [@owenlin0](https://github.com/owenlin0)\n\n- [#21556](https://github.com/openai/codex/pull/21556) codex-otel: add configurable trace metadata [@bbrown-oai](https://github.com/bbrown-oai)\n\n- [#21599](https://github.com/openai/codex/pull/21599) [codex] Apply a Dependabot cooldown of 7 days [@ww-oai](https://github.com/ww-oai)\n\n- [#21602](https://github.com/openai/codex/pull/21602) Use `--locked` in cargo build and lint invocations [@zanie-oai](https://github.com/zanie-oai)\n\n- [#20664](https://github.com/openai/codex/pull/20664) Add stdio exec-server client transport [@starr-openai](https://github.com/starr-openai)\n\n- [#21596](https://github.com/openai/codex/pull/21596) [codex] Remove remote thread store implementation [@wiltzius-openai](https://github.com/wiltzius-openai)\n\n- [#20665](https://github.com/openai/codex/pull/20665) Make environment providers own default selection [@starr-openai](https://github.com/starr-openai)\n\n- [#21143](https://github.com/openai/codex/pull/21143) Route view_image through selected environments [@starr-openai](https://github.com/starr-openai)\n\n- [#20666](https://github.com/openai/codex/pull/20666) Add CODEX_HOME environments TOML provider [@starr-openai](https://github.com/starr-openai)\n\n- [#21642](https://github.com/openai/codex/pull/21642) Send response.processed after remote compaction v2 [@pakrym-oai](https://github.com/pakrym-oai)\n\n- [#21646](https://github.com/openai/codex/pull/21646) Revert \"Use `--locked` in cargo build and lint invocations\" [@pakrym-oai](https://github.com/pakrym-oai)\n\n- [#21434](https://github.com/openai/codex/pull/21434) [codex-analytics] plumb protocol-native review timing [@rhan-oai](https://github.com/rhan-oai)\n\n- [#21485](https://github.com/openai/codex/pull/21485) Remove skills list extra roots [@xli-oai](https://github.com/xli-oai)\n\n- [#21623](https://github.com/openai/codex/pull/21623) feat: enable AWS login credentials for Bedrock auth [@celia-oai](https://github.com/celia-oai)\n\n- [#21637](https://github.com/openai/codex/pull/21637) feat: Update plugin share settings with discoverability [@xl-openai](https://github.com/xl-openai)\n\n- [#21685](https://github.com/openai/codex/pull/21685) Fix duplicate CLI issue template description [@etraut-openai](https://github.com/etraut-openai)\n\n- [#21686](https://github.com/openai/codex/pull/21686) Fix issue template labels [@etraut-openai](https://github.com/etraut-openai)\n\n- [#21688](https://github.com/openai/codex/pull/21688) Fix feature request Contributing link [@etraut-openai](https://github.com/etraut-openai)\n\n- [#21683](https://github.com/openai/codex/pull/21683) Remove exec research preview banner wording [@etraut-openai](https://github.com/etraut-openai)\n\n- [#21676](https://github.com/openai/codex/pull/21676) Omit service_tier from remote /responses/compact requests under API auth [@aibrahim-oai](https://github.com/aibrahim-oai)","publishedAt":"2026-05-08T00:00:00.000Z","url":"https://developers.openai.com/codex/changelog/#github-release-319777370","media":[],"prerelease":false,"source":{"slug":"openai-codex-changelog","name":"Codex Changelog","type":"feed"},"org":{"slug":"openai","name":"OpenAI"},"product":{"slug":"codex","name":"Codex"}},{"id":"rel_0gXWyD8xwhpW4jDQbdPJ6","version":"2.1.133","type":"feature","title":"2.1.133","summary":"- Added `worktree.baseRef` setting (`fresh` | `head`) to choose whether `--worktree`, `EnterWorktree`, and agent-isolation worktrees branch from `orig...","content":"- Added `worktree.baseRef` setting (`fresh` | `head`) to choose whether `--worktree`, `EnterWorktree`, and agent-isolation worktrees branch from `origin/` or local `HEAD`. **Note:** the default `fresh` changes `EnterWorktree`'s base back to `origin/` (it has been local `HEAD` since 2.1.128) — set `worktree.baseRef: \"head\"` to keep unpushed commits in new worktrees\n\n- Added `sandbox.bwrapPath` and `sandbox.socatPath` managed settings (Linux/WSL) to specify custom bubblewrap and socat binary locations\n\n- Added `parentSettingsBehavior` admin-tier key (`'first-wins' | 'merge'`) to let admins opt SDK `managedSettings` (parent tier) into the policy merge\n\n- Hooks now receive the active effort level via the `effort.level` JSON input field and the `$CLAUDE_EFFORT` environment variable, and Bash tool commands can read `$CLAUDE_EFFORT`\n\n- Improved focus mode behavior\n\n- Improved memory usage by releasing warm-spare background workers under memory pressure\n\n- Fixed parallel sessions all dead-ending at 401 after a refresh-token race wiped shared credentials\n\n- Fixed `Edit`/`Write` allow rules scoped to a drive root (`C:\\`) or POSIX `/` matching incorrectly and always prompting\n\n- Fixed an unhandled rejection (`ECOMPROMISED`) when a history or session-log file lock is compromised by clock skew or slow disk\n\n- Fixed pressing Esc during conversation compaction showing a spurious \"Error compacting conversation\" notification\n\n- Fixed `HTTP(S)_PROXY` / `NO_PROXY` / mTLS not being respected for the full MCP OAuth flow including discovery, dynamic client registration, token exchange, and token refresh\n\n- Fixed Read/Write/Edit being denied on mapped network drives passed via `--add-dir` / SDK `additionalDirectories`\n\n- Fixed Remote Control stop/interrupt from claude.ai not fully canceling the CLI session the same way local Esc does, causing queued messages to never advance after interrupting a stuck tool or prompt\n\n- Fixed `/effort` in one session unexpectedly changing the effort level of other concurrent sessions, and a related issue where an IDE effort change could be silently dropped\n\n- Fixed subagents not discovering project, user, or plugin skills via the Skill tool\n\n- `claude --help` now lists `--remote-control` alongside `--remote-control-session-name-prefix`\n\n- [VSCode] Fixed `claudeCode.claudeProcessWrapper` failing with \"Unsupported platform\" when the extension build doesn't bundle a Claude binary","publishedAt":"2026-05-07T23:53:32.000Z","url":"https://code.claude.com/docs/en/changelog#2-1-133","media":[],"prerelease":false,"source":{"slug":"claude-code","name":"Claude Code","type":"scrape"},"org":{"slug":"anthropic","name":"Anthropic"},"product":{"slug":"claude-code","name":"Claude Code"}},{"id":"rel_sAFuUe_8c1F39tu8gXSmS","version":"v3.35.0","type":"feature","title":"v3.35.0","summary":"## 3.35.0 (2026-05-07)\n\nFull Changelog: [v3.34.0...v3.35.0](https://github.com/openai/openai-go/compare/v3.34.0...v3.35.0)\n\n### Features\n\n* **api:** a...","content":"## 3.35.0 (2026-05-07)\n\nFull Changelog: [v3.34.0...v3.35.0](https://github.com/openai/openai-go/compare/v3.34.0...v3.35.0)\n\n### Features\n\n* **api:** add quantity field to admin organization usage costs results ([899b6de](https://github.com/openai/openai-go/commit/899b6dec70bd180a98497150b892285261882625))\n* **api:** launch realtime translate + update image 2 ([1fbf3a6](https://github.com/openai/openai-go/commit/1fbf3a68e418f84993db1c291255c467e47dbe37))\n* **api:** manual updates ([cd9047f](https://github.com/openai/openai-go/commit/cd9047f7416cedba79c7479d96acce3f59ac2c86))\n* **api:** realtime 2 ([5b51410](https://github.com/openai/openai-go/commit/5b514101d40d0b9f36194bab0c7e49a7f6098c03))\n\n\n### Bug Fixes\n\n* **api:** fix imagegen `size` enum regression ([995778e](https://github.com/openai/openai-go/commit/995778eac4b36433a7d116f9a09999f34e78b91f))\n* **types:** add format tags for URI and timestamp fields across models ([a6c2198](https://github.com/openai/openai-go/commit/a6c219806da98fa140b837a1063d3ee7f5f51e8d))\n\n\n### Chores\n\n* **internal:** regenerate SDK with no functional changes ([71c80ef](https://github.com/openai/openai-go/commit/71c80ef9e4407d6714f9d0013bc9579dfab1a223))\n* redact api-key headers in debug logs ([ff51a02](https://github.com/openai/openai-go/commit/ff51a02d264f06535bac6000865b8ad27e4e8a72))\n\n\n### Documentation\n\n* **api:** document web_search_call.results output data option ([6c2d8ef](https://github.com/openai/openai-go/commit/6c2d8eff506aae6b4fea17fdca94748bd4b2cd3e))","publishedAt":"2026-05-07T17:33:38.000Z","url":"https://github.com/openai/openai-go/releases/tag/v3.35.0","media":[],"prerelease":false,"source":{"slug":"openai-go-sdk","name":"Go SDK","type":"github"},"org":{"slug":"openai","name":"OpenAI"},"product":null},{"id":"rel_e5fGrhftkd4O2xxJbcVYy","version":"v6.37.0","type":"feature","title":"v6.37.0","summary":"## 6.37.0 (2026-05-07)\n\nFull Changelog: [v6.36.0...v6.37.0](https://github.com/openai/openai-node/compare/v6.36.0...v6.37.0)\n\n### Features\n\n* **api:**...","content":"## 6.37.0 (2026-05-07)\n\nFull Changelog: [v6.36.0...v6.37.0](https://github.com/openai/openai-node/compare/v6.36.0...v6.37.0)\n\n### Features\n\n* **api:** add quantity field to admin organization usage responses ([273a8f7](https://github.com/openai/openai-node/commit/273a8f73e65839fb4d2ebc2a54ccc38383bd7517))\n* **api:** add web_search_call.results output option to responses ([91c75e0](https://github.com/openai/openai-node/commit/91c75e064f12946d3774cb182e5cd795000019ce))\n* **api:** launch realtime translate + update image 2 ([a296b66](https://github.com/openai/openai-node/commit/a296b661f0598ed001309d7deb693caa1596875c))\n* **api:** manual updates ([794b905](https://github.com/openai/openai-node/commit/794b90513f5ca63bd95c6cf774722fa6aeae293d))\n* **api:** manual updates ([6963729](https://github.com/openai/openai-node/commit/69637293939580a9d20bd9274d86f23a012ffd87))\n* **api:** realtime 2 ([f4b7177](https://github.com/openai/openai-node/commit/f4b717702182adfc684270c0e7d5ff11d6c7a408))\n\n\n### Bug Fixes\n\n* **api:** fix imagegen `size` enum regression ([4fe8469](https://github.com/openai/openai-node/commit/4fe8469a7c1f892d586df375a449f6bd0e318b15))\n\n\n### Chores\n\n* redact api-key headers in debug logs ([99c9c80](https://github.com/openai/openai-node/commit/99c9c800e179c13659b96386dedd6c17dd6ffb8d))","publishedAt":"2026-05-07T17:33:30.000Z","url":"https://github.com/openai/openai-node/releases/tag/v6.37.0","media":[],"prerelease":false,"source":{"slug":"openai-node-sdk","name":"Node SDK","type":"github"},"org":{"slug":"openai","name":"OpenAI"},"product":null},{"id":"rel_QfljMni2BEBiJYVIGvnqM","version":"v2.36.0","type":"feature","title":"v2.36.0","summary":"## 2.36.0 (2026-05-07)\n\nFull Changelog: [v2.35.1...v2.36.0](https://github.com/openai/openai-python/compare/v2.35.1...v2.36.0)\n\n### Features\n\n* **api:...","content":"## 2.36.0 (2026-05-07)\n\nFull Changelog: [v2.35.1...v2.36.0](https://github.com/openai/openai-python/compare/v2.35.1...v2.36.0)\n\n### Features\n\n* **api:** manual updates ([13c639c](https://github.com/openai/openai-python/commit/13c639cc7d57e4fbd4406563511e15eeb88a54b2))\n* **api:** realtime 2 ([8fe0ab8](https://github.com/openai/openai-python/commit/8fe0ab87e67eeb3cc27426b50093845229520f0e))","publishedAt":"2026-05-07T17:33:02.000Z","url":"https://github.com/openai/openai-python/releases/tag/v2.36.0","media":[],"prerelease":false,"source":{"slug":"openai-python-sdk","name":"Python SDK","type":"github"},"org":{"slug":"openai","name":"OpenAI"},"product":null},{"id":"rel_aoKhHOym3kXGwsrWQ6CFH","version":null,"type":"feature","title":"Advancing voice intelligence with new models in the API","summary":"Explore new realtime voice models in the OpenAI API that can reason, translate, and transcribe speech, enabling more natural and intelligent voice exp...","content":"Explore new realtime voice models in the OpenAI API that can reason, translate, and transcribe speech, enabling more natural and intelligent voice experiences.","publishedAt":"2026-05-07T10:00:00.000Z","url":"https://openai.com/index/advancing-voice-intelligence-with-new-models-in-the-api","media":[],"prerelease":false,"source":{"slug":"openai-news","name":"Product Releases","type":"feed"},"org":{"slug":"openai","name":"OpenAI"},"product":null},{"id":"rel_fmUdrdBdGTYvfSSWIiv8X","version":null,"type":"feature","title":"Codex for Chrome","summary":"# Codex for Chrome\n\nWith the new extension for Chrome, Codex is even better at working with apps\nand websites in your browser. It works in parallel ac...","content":"# Codex for Chrome\n\nWith the new extension for Chrome, Codex is even better at working with apps\nand websites in your browser. It works in parallel across tabs in the\nbackground without taking over your browser, and you stay in control of which\nwebsites Codex can use.\n\nLearn more in the [Codex Chrome extension documentation](/codex/app/chrome-extension).","publishedAt":"2026-05-07T00:00:00.000Z","url":"https://developers.openai.com/codex/changelog/#codex-2026-05-07","media":[],"prerelease":false,"source":{"slug":"openai-codex-changelog","name":"Codex Changelog","type":"feed"},"org":{"slug":"openai","name":"OpenAI"},"product":{"slug":"codex","name":"Codex"}},{"id":"rel_PaxNM-2TTtPJoRwoVMRBd","version":"0.129.0","type":"feature","title":"Codex CLI Release: 0.129.0","summary":"## New Features\n\n- The TUI now supports modal Vim editing in the composer, including `/vim`, default-mode config, and Vim-specific keymap contexts. ([...","content":"## New Features\n\n- The TUI now supports modal Vim editing in the composer, including `/vim`, default-mode config, and Vim-specific keymap contexts. ([#18595](https://github.com/openai/codex/pull/18595))\n\n- TUI workflows are easier to resume and copy from with a redesigned resume/fork picker, raw scrollback mode, `/ide` context injection, and workspace-aware `/diff`. ([#20065](https://github.com/openai/codex/pull/20065), [#20819](https://github.com/openai/codex/pull/20819), [#20294](https://github.com/openai/codex/pull/20294), [#21001](https://github.com/openai/codex/pull/21001))\n\n- The status line can show theme-aware colors plus optional PR and branch-change summaries, and `/keymap debug` helps inspect terminal key events. ([#19631](https://github.com/openai/codex/pull/19631), [#20892](https://github.com/openai/codex/pull/20892), [#20794](https://github.com/openai/codex/pull/20794))\n\n- Plugin management now supports workspace sharing, share access controls, source filtering, local share path tracking, marketplace removal/upgrades, remote bundle sync, and admin-disabled status handling. ([#20278](https://github.com/openai/codex/pull/20278), [#21124](https://github.com/openai/codex/pull/21124), [#21419](https://github.com/openai/codex/pull/21419), [#20560](https://github.com/openai/codex/pull/20560), [#19843](https://github.com/openai/codex/pull/19843), [#20478](https://github.com/openai/codex/pull/20478), [#20268](https://github.com/openai/codex/pull/20268), [#20298](https://github.com/openai/codex/pull/20298))\n\n- Hooks can be browsed and toggled from `/hooks`, can run before/after compaction, and can add `PreToolUse` context; Codex Apps auth and eligible MCP elicitations now surface through TUI/Guardian flows. ([#19882](https://github.com/openai/codex/pull/19882), [#19905](https://github.com/openai/codex/pull/19905), [#20692](https://github.com/openai/codex/pull/20692), [#19193](https://github.com/openai/codex/pull/19193), [#19431](https://github.com/openai/codex/pull/19431))\n\n- Experimental goals are now discoverable, stay paused across resume unless the user opts back in, and show clearer validation and multi-day duration output. ([#20083](https://github.com/openai/codex/pull/20083), [#20790](https://github.com/openai/codex/pull/20790), [#20746](https://github.com/openai/codex/pull/20746), [#20558](https://github.com/openai/codex/pull/20558))\n\n## Bug Fixes\n\n- `/copy` works better in tmux, Alt+Enter and modified Delete/Backspace keys behave correctly, and Windows typing/paste latency was reduced. ([#20207](https://github.com/openai/codex/pull/20207), [#20535](https://github.com/openai/codex/pull/20535), [#21058](https://github.com/openai/codex/pull/21058), [#18914](https://github.com/openai/codex/pull/18914))\n\n- Large paste placeholders and Ctrl+C-stashed drafts now survive clear/editor workflows without corrupting draft history. ([#21091](https://github.com/openai/codex/pull/21091), [#21190](https://github.com/openai/codex/pull/21190), [#21351](https://github.com/openai/codex/pull/21351), [#21397](https://github.com/openai/codex/pull/21397))\n\n- TUI startup and accessibility were tightened by bounding terminal probes, clearing the first inline viewport render, and honoring `animations = false` for live rows. ([#20654](https://github.com/openai/codex/pull/20654), [#21450](https://github.com/openai/codex/pull/21450), [#20564](https://github.com/openai/codex/pull/20564))\n\n- Linux sandbox startup is more reliable across older `bwrap`, slow mount probes, symlink-protected paths, and shared `/tmp` setups. ([#20628](https://github.com/openai/codex/pull/20628), [#20111](https://github.com/openai/codex/pull/20111), [#21127](https://github.com/openai/codex/pull/21127), [#21234](https://github.com/openai/codex/pull/21234))\n\n- Windows sandbox and exec policy now handle named pipes, ConPTY teardown, PowerShell-wrapped allow rules, worktree `safe.directory`, and unsafe Git options more reliably. ([#20270](https://github.com/openai/codex/pull/20270), [#20685](https://github.com/openai/codex/pull/20685), [#20336](https://github.com/openai/codex/pull/20336), [#21409](https://github.com/openai/codex/pull/21409), [#21275](https://github.com/openai/codex/pull/21275))\n\n- Fixed custom CA login behind TLS-inspecting proxies, Bedrock runtime endpoint reporting, dangerous project config keys, heredoc redirect approval matching, and unbounded MCP/hook output growth. ([#20676](https://github.com/openai/codex/pull/20676), [#20275](https://github.com/openai/codex/pull/20275), [#20098](https://github.com/openai/codex/pull/20098), [#20113](https://github.com/openai/codex/pull/20113), [#20260](https://github.com/openai/codex/pull/20260), [#21069](https://github.com/openai/codex/pull/21069))\n\n## Documentation\n\n- Updated the embedded OpenAI Docs sample skill so API-key setup guidance stays aligned with other docs variants. ([#21263](https://github.com/openai/codex/pull/21263))\n\n- Documented how generated git commit attribution is gated by `codex_git_commit` and configured in `config.toml`. ([#21379](https://github.com/openai/codex/pull/21379))\n\n- Removed local-only planning/spec docs and redirected config docs toward the maintained external documentation surface. ([#20896](https://github.com/openai/codex/pull/20896))\n\n## Chores\n\n- Linux releases now build, publish, bundle, and verify a standalone `bwrap` fallback for npm and DotSlash installs. ([#21255](https://github.com/openai/codex/pull/21255), [#21256](https://github.com/openai/codex/pull/21256), [#21257](https://github.com/openai/codex/pull/21257), [#21312](https://github.com/openai/codex/pull/21312), [#21285](https://github.com/openai/codex/pull/21285))\n\n- Vendored Bubblewrap was updated to 0.11.2, including upstream security changes around setuid support. ([#21389](https://github.com/openai/codex/pull/21389))\n\n- Windows Bazel CI now uses faster cross-compilation for tests, clippy, and release-build checks, and Bazel now runs sharded Rust integration tests. ([#20585](https://github.com/openai/codex/pull/20585), [#20701](https://github.com/openai/codex/pull/20701), [#21057](https://github.com/openai/codex/pull/21057))\n\n- App-server and protocol internals were split and slimmed down, including transport extraction, protocol module decomposition, thread/message history moves, and tool-handler cleanup. ([#20324](https://github.com/openai/codex/pull/20324), [#20325](https://github.com/openai/codex/pull/20325), [#20348](https://github.com/openai/codex/pull/20348), [#20545](https://github.com/openai/codex/pull/20545), [#21251](https://github.com/openai/codex/pull/21251), [#21278](https://github.com/openai/codex/pull/21278), [#21395](https://github.com/openai/codex/pull/21395))\n\n- Analytics and diagnostics coverage expanded for tool lifecycles, goals, plugin skills, thread sources, service tiers, and PR issue labeling. ([#17089](https://github.com/openai/codex/pull/17089), [#17090](https://github.com/openai/codex/pull/17090), [#20799](https://github.com/openai/codex/pull/20799), [#20923](https://github.com/openai/codex/pull/20923), [#20949](https://github.com/openai/codex/pull/20949), [#20969](https://github.com/openai/codex/pull/20969), [#20893](https://github.com/openai/codex/pull/20893))\n\n## Changelog\n\nFull Changelog: [rust-v0.128.0...rust-v0.129.0](https://github.com/openai/codex/compare/rust-v0.128.0...rust-v0.129.0)\n\n- [#20278](https://github.com/openai/codex/pull/20278) feat: Add workspace plugin sharing APIs [@xl-openai](https://github.com/xl-openai)\n\n- [#20334](https://github.com/openai/codex/pull/20334) Make missing config clears no-ops [@etraut-openai](https://github.com/etraut-openai)\n\n- [#20246](https://github.com/openai/codex/pull/20246) Gate multi-agent v2 tools independently of collab [@jif-oai](https://github.com/jif-oai)\n\n- [#20361](https://github.com/openai/codex/pull/20361) realtime: rename provider session ids [@aibrahim-oai](https://github.com/aibrahim-oai)\n\n- [#20260](https://github.com/openai/codex/pull/20260) fix(core): truncate large mcp tool outputs in rollouts [@owenlin0](https://github.com/owenlin0)\n\n- [#20083](https://github.com/openai/codex/pull/20083) Mark goals feature as experimental [@etraut-openai](https://github.com/etraut-openai)\n\n- [#19843](https://github.com/openai/codex/pull/19843) /plugins: remove marketplace [@canvrno-oai](https://github.com/canvrno-oai)\n\n- [#20458](https://github.com/openai/codex/pull/20458) [Extension] Allowlist Chrome Extension in the tool_suggest tool [@teddywyly-oai](https://github.com/teddywyly-oai)\n\n- [#20324](https://github.com/openai/codex/pull/20324) Remove core protocol dependency [1/2] [@etraut-openai](https://github.com/etraut-openai)\n\n- [#20299](https://github.com/openai/codex/pull/20299) Move item event mapping into app-server-protocol [@pakrym-oai](https://github.com/pakrym-oai)\n\n- [#20325](https://github.com/openai/codex/pull/20325) Remove core protocol dependency [2/2] [@etraut-openai](https://github.com/etraut-openai)\n\n- [#20471](https://github.com/openai/codex/pull/20471) Stop emitting item/fileChange/outputDelta output delta notifications [@pakrym-oai](https://github.com/pakrym-oai)\n\n- [#20245](https://github.com/openai/codex/pull/20245) [Codex] Add browser use external feature flag [@khoi-oai](https://github.com/khoi-oai)\n\n- [#19882](https://github.com/openai/codex/pull/19882) Add /hooks browser for lifecycle hooks [@abhinav-oai](https://github.com/abhinav-oai)\n\n- [#20275](https://github.com/openai/codex/pull/20275) fix: show correct Bedrock runtime endpoint in /status [@celia-oai](https://github.com/celia-oai)\n\n- [#20270](https://github.com/openai/codex/pull/20270) [codex] Fix elevated Windows sandbox named-pipe access [@iceweasel-oai](https://github.com/iceweasel-oai)\n\n- [#20463](https://github.com/openai/codex/pull/20463) feat(rollouts): store EventMsg::ApplyPatchEnd in limited history mode [@owenlin0](https://github.com/owenlin0)\n\n- [#20101](https://github.com/openai/codex/pull/20101) install WFP filters for Windows sandbox setup [@iceweasel-oai](https://github.com/iceweasel-oai)\n\n- [#20474](https://github.com/openai/codex/pull/20474) [plugin] Add Canva to suggesteable list. [@mzeng-openai](https://github.com/mzeng-openai)\n\n- [#20379](https://github.com/openai/codex/pull/20379) Send external import completion for sync imports [@alexsong-oai](https://github.com/alexsong-oai)\n\n- [#19280](https://github.com/openai/codex/pull/19280) [codex] Migrate thread turns list to thread store [@wiltzius-openai](https://github.com/wiltzius-openai)\n\n- [#20348](https://github.com/openai/codex/pull/20348) Move plugin out of core. [@xl-openai](https://github.com/xl-openai)\n\n- [#19160](https://github.com/openai/codex/pull/19160) Make apply_patch streaming parser stateful [@akshaynathan](https://github.com/akshaynathan)\n\n- [#20504](https://github.com/openai/codex/pull/20504) fix flaky test falls_back_to_registered_fallback_port_when_default_po… [@owenlin0](https://github.com/owenlin0)\n\n- [#20098](https://github.com/openai/codex/pull/20098) fix: ignore dangerous project-level config keys [@owenlin0](https://github.com/owenlin0)\n\n- [#20268](https://github.com/openai/codex/pull/20268) Sync remote installed plugin bundles [@xli-oai](https://github.com/xli-oai)\n\n- [#20502](https://github.com/openai/codex/pull/20502) fix(tui): set persist_extended_history: false [@owenlin0](https://github.com/owenlin0)\n\n- [#20069](https://github.com/openai/codex/pull/20069) Bypass review for always-allow MCP tools in auto-review [@maja-openai](https://github.com/maja-openai)\n\n- [#18595](https://github.com/openai/codex/pull/18595) feat(tui): add vim composer mode [@fcoury-oai](https://github.com/fcoury-oai)\n\n- [#20267](https://github.com/openai/codex/pull/20267) Emit analytics for remote plugin installs [@xli-oai](https://github.com/xli-oai)\n\n- [#20499](https://github.com/openai/codex/pull/20499) fix(app-server): mark thread/turns/list and exclude_turns as experime… [@owenlin0](https://github.com/owenlin0)\n\n- [#20522](https://github.com/openai/codex/pull/20522) Alias codex_hooks feature as hooks [@abhinav-oai](https://github.com/abhinav-oai)\n\n- [#20336](https://github.com/openai/codex/pull/20336) execpolicy: unwrap PowerShell -Command wrappers on Windows [@iceweasel-oai](https://github.com/iceweasel-oai)\n\n- [#20113](https://github.com/openai/codex/pull/20113) fix(exec_policy) heredoc parsing file_redirect [@dylan-hurd-oai](https://github.com/dylan-hurd-oai)\n\n- [#20341](https://github.com/openai/codex/pull/20341) app-server: switch remote control to protocol v3 segmentation [@euroelessar](https://github.com/euroelessar)\n\n- [#20300](https://github.com/openai/codex/pull/20300) [codex-analytics] centralize thread analytics state [@rhan-oai](https://github.com/rhan-oai)\n\n- [#20484](https://github.com/openai/codex/pull/20484) [codex] Improve PR babysitter CI diagnostics and guardrails [@wiltzius-openai](https://github.com/wiltzius-openai)\n\n- [#20298](https://github.com/openai/codex/pull/20298) Surface admin-disabled remote plugin status [@xli-oai](https://github.com/xli-oai)\n\n- [#20511](https://github.com/openai/codex/pull/20511) [codex] Remove unused event messages [@pakrym-oai](https://github.com/pakrym-oai)\n\n- [#19474](https://github.com/openai/codex/pull/19474) Make thread store process-scoped [@wiltzius-openai](https://github.com/wiltzius-openai)\n\n- [#20558](https://github.com/openai/codex/pull/20558) Format multi-day goal durations in the TUI [@etraut-openai](https://github.com/etraut-openai)\n\n- [#19631](https://github.com/openai/codex/pull/19631) Color TUI statusline from active theme [@etraut-openai](https://github.com/etraut-openai)\n\n- [#20265](https://github.com/openai/codex/pull/20265) Refresh remote plugin cache on auth changes [@xli-oai](https://github.com/xli-oai)\n\n- [#20150](https://github.com/openai/codex/pull/20150) Add remote plugin skill read API [@xli-oai](https://github.com/xli-oai)\n\n- [#20560](https://github.com/openai/codex/pull/20560) feat: Track local paths for shared plugins [@xl-openai](https://github.com/xl-openai)\n\n- [#20600](https://github.com/openai/codex/pull/20600) chore: allow memories edition [@jif-oai](https://github.com/jif-oai)\n\n- [#20602](https://github.com/openai/codex/pull/20602) feat: ad-hoc instructions [@jif-oai](https://github.com/jif-oai)\n\n- [#20610](https://github.com/openai/codex/pull/20610) chore: improve remember prompt [@jif-oai](https://github.com/jif-oai)\n\n- [#20606](https://github.com/openai/codex/pull/20606) feat: seed ad-hoc memory extension instructions [@jif-oai](https://github.com/jif-oai)\n\n- [#20405](https://github.com/openai/codex/pull/20405) feat: export and replay effective config locks [@jif-oai](https://github.com/jif-oai)\n\n- [#20540](https://github.com/openai/codex/pull/20540) Move apply-patch file changes into turn items [@pakrym-oai](https://github.com/pakrym-oai)\n\n- [#20564](https://github.com/openai/codex/pull/20564) Enforce `animations = false` for screen readers [@etraut-openai](https://github.com/etraut-openai)\n\n- [#20523](https://github.com/openai/codex/pull/20523) Remove no-tool goal continuation suppression [@etraut-openai](https://github.com/etraut-openai)\n\n- [#20627](https://github.com/openai/codex/pull/20627) fix: cargo deny [@jif-oai](https://github.com/jif-oai)\n\n- [#20545](https://github.com/openai/codex/pull/20545) app-server: move transport into dedicated crate [@euroelessar](https://github.com/euroelessar)\n\n- [#20294](https://github.com/openai/codex/pull/20294) Add /ide context support to the TUI [@etraut-openai](https://github.com/etraut-openai)\n\n- [#20630](https://github.com/openai/codex/pull/20630) [codex] Add Codex environment config [@pakrym-oai](https://github.com/pakrym-oai)\n\n- [#20524](https://github.com/openai/codex/pull/20524) deprecate legacy notify [@abhinav-oai](https://github.com/abhinav-oai)\n\n- [#20486](https://github.com/openai/codex/pull/20486) [codex] Migrate loaded thread/read history to ThreadStore [@wiltzius-openai](https://github.com/wiltzius-openai)\n\n- [#20281](https://github.com/openai/codex/pull/20281) Use selected turn environments for runtime context [@starr-openai](https://github.com/starr-openai)\n\n- [#20535](https://github.com/openai/codex/pull/20535) fix(tui): restore alt-enter newline alias [@fcoury-oai](https://github.com/fcoury-oai)\n\n- [#20650](https://github.com/openai/codex/pull/20650) fix: reduce ConfigBuilder::build stack usage [@jif-oai](https://github.com/jif-oai)\n\n- [#20478](https://github.com/openai/codex/pull/20478) /plugins: add marketplace upgrade flow [@canvrno-oai](https://github.com/canvrno-oai)\n\n- [#20512](https://github.com/openai/codex/pull/20512) [codex] Emit image view as core item [@pakrym-oai](https://github.com/pakrym-oai)\n\n- [#20562](https://github.com/openai/codex/pull/20562) Use the 2025-06-18 elicitation capability shape [@abhinav-oai](https://github.com/abhinav-oai)\n\n- [#20674](https://github.com/openai/codex/pull/20674) Clear live hook rows when turns finalize [@abhinav-oai](https://github.com/abhinav-oai)\n\n- [#20646](https://github.com/openai/codex/pull/20646) Surface multi-environment choices in environment context [@starr-openai](https://github.com/starr-openai)\n\n- [#20542](https://github.com/openai/codex/pull/20542) Prune unused code-mode globals [@cconger](https://github.com/cconger)\n\n- [#20585](https://github.com/openai/codex/pull/20585) ci: cross-compile Windows Bazel tests [@bolinfest](https://github.com/bolinfest)\n\n- [#20701](https://github.com/openai/codex/pull/20701) ci: cross-compile Windows Bazel clippy [@bolinfest](https://github.com/bolinfest)\n\n- [#20676](https://github.com/openai/codex/pull/20676) Fix custom CA login behind TLS-inspecting proxies [@jgershen-oai](https://github.com/jgershen-oai)\n\n- [#20654](https://github.com/openai/codex/pull/20654) fix(tui): bound startup terminal probes [@fcoury-oai](https://github.com/fcoury-oai)\n\n- [#20566](https://github.com/openai/codex/pull/20566) [tool_suggest] More prompt polishes. [@mzeng-openai](https://github.com/mzeng-openai)\n\n- [#20751](https://github.com/openai/codex/pull/20751) Bound websocket request sends with idle timeout [@pakrym-oai](https://github.com/pakrym-oai)\n\n- [#20893](https://github.com/openai/codex/pull/20893) [codex] Add issue labeler area labels [@etraut-openai](https://github.com/etraut-openai)\n\n- [#20896](https://github.com/openai/codex/pull/20896) Remove local docs and specs [@etraut-openai](https://github.com/etraut-openai)\n\n- [#20897](https://github.com/openai/codex/pull/20897) [codex] Refactor app-server dispatch result flow [@pakrym-oai](https://github.com/pakrym-oai)\n\n- [#20677](https://github.com/openai/codex/pull/20677) [codex] Emit MCP tool calls as turn items [@pakrym-oai](https://github.com/pakrym-oai)\n\n- [#20973](https://github.com/openai/codex/pull/20973) feat: support template interpolation in multi-agent usage hints [@jif-oai](https://github.com/jif-oai)\n\n- [#20622](https://github.com/openai/codex/pull/20622) feat: memories mcp v1 [@jif-oai](https://github.com/jif-oai)\n\n- [#20773](https://github.com/openai/codex/pull/20773) feat: add remote compaction v2 Responses client path [@jif-oai](https://github.com/jif-oai)\n\n- [#20986](https://github.com/openai/codex/pull/20986) feat: add line offsets to memory read MCP [@jif-oai](https://github.com/jif-oai)\n\n- [#20991](https://github.com/openai/codex/pull/20991) feat: add max_lines to memories MCP read [@jif-oai](https://github.com/jif-oai)\n\n- [#20993](https://github.com/openai/codex/pull/20993) feat: paginate MCP memories list [@jif-oai](https://github.com/jif-oai)\n\n- [#20994](https://github.com/openai/codex/pull/20994) feat: make memories MCP list shallow [@jif-oai](https://github.com/jif-oai)\n\n- [#20996](https://github.com/openai/codex/pull/20996) feat: paginate memories MCP search results [@jif-oai](https://github.com/jif-oai)\n\n- [#20997](https://github.com/openai/codex/pull/20997) feat: add context lines to memories MCP search [@jif-oai](https://github.com/jif-oai)\n\n- [#20998](https://github.com/openai/codex/pull/20998) nit: renaming [@jif-oai](https://github.com/jif-oai)\n\n- [#21004](https://github.com/openai/codex/pull/21004) feat: support multi-query memories search [@jif-oai](https://github.com/jif-oai)\n\n- [#21006](https://github.com/openai/codex/pull/21006) nit: legacy [@jif-oai](https://github.com/jif-oai)\n\n- [#20815](https://github.com/openai/codex/pull/20815) Speed up /side parent restore replay [@etraut-openai](https://github.com/etraut-openai)\n\n- [#20790](https://github.com/openai/codex/pull/20790) Keep paused goals paused on thread resume [@etraut-openai](https://github.com/etraut-openai)\n\n- [#20940](https://github.com/openai/codex/pull/20940) [codex] Split app-server request processors [@pakrym-oai](https://github.com/pakrym-oai)\n\n- [#21023](https://github.com/openai/codex/pull/21023) typo [@jif-oai](https://github.com/jif-oai)\n\n- [#21012](https://github.com/openai/codex/pull/21012) memories/mcp: generate tool schemas with schemars [@jif-oai](https://github.com/jif-oai)\n\n- [#21010](https://github.com/openai/codex/pull/21010) memories-mcp: reject symlink traversal in local backend [@jif-oai](https://github.com/jif-oai)\n\n- [#20989](https://github.com/openai/codex/pull/20989) core: share responses request builder with compact requests [@jif-oai](https://github.com/jif-oai)\n\n- [#20853](https://github.com/openai/codex/pull/20853) [mcp-apps] Persist MCP Apps specific tool call end event. [@mzeng-openai](https://github.com/mzeng-openai)\n\n- [#20750](https://github.com/openai/codex/pull/20750) Unify skip-review handling for approval_mode = \"approve\" [@mzeng-openai](https://github.com/mzeng-openai)\n\n- [#20682](https://github.com/openai/codex/pull/20682) feat(app-server): always return limited thread history [@owenlin0](https://github.com/owenlin0)\n\n- [#20628](https://github.com/openai/codex/pull/20628) fix(linux-sandbox): fall back when system bwrap lacks perms [@viyatb-oai](https://github.com/viyatb-oai)\n\n- [#20794](https://github.com/openai/codex/pull/20794) feat(tui): add keymap debug inspector [@fcoury-oai](https://github.com/fcoury-oai)\n\n- [#21034](https://github.com/openai/codex/pull/21034) tui: retire /approvals and rename /autoreview to /approve [@won-openai](https://github.com/won-openai)\n\n- [#20669](https://github.com/openai/codex/pull/20669) Prepare selected environment plumbing [@starr-openai](https://github.com/starr-openai)\n\n- [#20685](https://github.com/openai/codex/pull/20685) Fix Windows PTY teardown by preserving ConPTY ownership [@iceweasel-oai](https://github.com/iceweasel-oai)\n\n- [#20663](https://github.com/openai/codex/pull/20663) Add stdio exec-server listener [@starr-openai](https://github.com/starr-openai)\n\n- [#20561](https://github.com/openai/codex/pull/20561) state: pass state db handles through consumers [@euroelessar](https://github.com/euroelessar)\n\n- [#21054](https://github.com/openai/codex/pull/21054) rollout: store web search and mcp tool calls [@owenlin0](https://github.com/owenlin0)\n\n- [#20892](https://github.com/openai/codex/pull/20892) feat(tui): add PR summary statusline items [@fcoury-oai](https://github.com/fcoury-oai)\n\n- [#20798](https://github.com/openai/codex/pull/20798) feat(tui): improve TUI keymap coverage [@fcoury-oai](https://github.com/fcoury-oai)\n\n- [#21053](https://github.com/openai/codex/pull/21053) Use MCP server instructions in deferred namespace descriptions [@sayan-oai](https://github.com/sayan-oai)\n\n- [#21026](https://github.com/openai/codex/pull/21026) core: preserve last model ids in feedback tags [@sayan-oai](https://github.com/sayan-oai)\n\n- [#21060](https://github.com/openai/codex/pull/21060) core: fix apply_patch request permissions test [@bolinfest](https://github.com/bolinfest)\n\n- [#20060](https://github.com/openai/codex/pull/20060) Add reasoning effort to turn tracing spans [@charley-openai](https://github.com/charley-openai)\n\n- [#21058](https://github.com/openai/codex/pull/21058) fix(tui): support modified backspace/delete keys [@fcoury-oai](https://github.com/fcoury-oai)\n\n- [#21057](https://github.com/openai/codex/pull/21057) bazel: run sharded rust integration tests [@bolinfest](https://github.com/bolinfest)\n\n- [#18914](https://github.com/openai/codex/pull/18914) fix(tui): use shared paste burst interval on Windows [@fcoury-oai](https://github.com/fcoury-oai)\n\n- [#20715](https://github.com/openai/codex/pull/20715) Make realtime sideband startup async [@kmeelu-oai](https://github.com/kmeelu-oai)\n\n- [#20514](https://github.com/openai/codex/pull/20514) [codex-analytics] add item lifecycle timing [@rhan-oai](https://github.com/rhan-oai)\n\n- [#20722](https://github.com/openai/codex/pull/20722) Remove remote plugin uninstall prefix gate [@xli-oai](https://github.com/xli-oai)\n\n- [#19040](https://github.com/openai/codex/pull/19040) [codex] Add unsandboxed process exec API [@euroelessar](https://github.com/euroelessar)\n\n- [#21105](https://github.com/openai/codex/pull/21105) [network-proxy] Cover DNS timeout blocking [@evawong-oai](https://github.com/evawong-oai)\n\n- [#21059](https://github.com/openai/codex/pull/21059) Rename agent identity login surface to access token [@shijie-oai](https://github.com/shijie-oai)\n\n- [#20576](https://github.com/openai/codex/pull/20576) codex: route metadata updates through ThreadStore [@wiltzius-openai](https://github.com/wiltzius-openai)\n\n- [#20923](https://github.com/openai/codex/pull/20923) Add plugin ID to skill analytics [@alexsong-oai](https://github.com/alexsong-oai)\n\n- [#21122](https://github.com/openai/codex/pull/21122) Add turn_id to Codex skill invocation analytics [@edwardysun3](https://github.com/edwardysun3)\n\n- [#20575](https://github.com/openai/codex/pull/20575) codex: migrate (more) app-server thread history reads to ThreadStore [@wiltzius-openai](https://github.com/wiltzius-openai)\n\n- [#21069](https://github.com/openai/codex/pull/21069) Spill large hook outputs from context [@abhinav-oai](https://github.com/abhinav-oai)\n\n- [#20969](https://github.com/openai/codex/pull/20969) 1- Add model service tiers metadata [@aibrahim-oai](https://github.com/aibrahim-oai)\n\n- [#21170](https://github.com/openai/codex/pull/21170) tools: remove unused experimental `list_dir` tool [@jif-oai](https://github.com/jif-oai)\n\n- [#21201](https://github.com/openai/codex/pull/21201) memories-mcp: hide dot paths from list, read, and search [@jif-oai](https://github.com/jif-oai)\n\n- [#21204](https://github.com/openai/codex/pull/21204) feat: support windowed multi-query memory search [@jif-oai](https://github.com/jif-oai)\n\n- [#21205](https://github.com/openai/codex/pull/21205) feat: add normalized matching to memory search [@jif-oai](https://github.com/jif-oai)\n\n- [#20207](https://github.com/openai/codex/pull/20207) fix(tui): make /copy work inside tmux without passthrough [@fcoury-oai](https://github.com/fcoury-oai)\n\n- [#20799](https://github.com/openai/codex/pull/20799) Add goal lifecycle metrics [@etraut-openai](https://github.com/etraut-openai)\n\n- [#20746](https://github.com/openai/codex/pull/20746) Validate /goal objective length in TUI [@etraut-openai](https://github.com/etraut-openai)\n\n- [#20708](https://github.com/openai/codex/pull/20708) Add Windows sandbox readiness RPC [@iceweasel-oai](https://github.com/iceweasel-oai)\n\n- [#20692](https://github.com/openai/codex/pull/20692) Support PreToolUse additionalContext [@abhinav-oai](https://github.com/abhinav-oai)\n\n- [#21091](https://github.com/openai/codex/pull/21091) [codex] Fix TUI large paste placeholder numbering after Ctrl+C [@canvrno-oai](https://github.com/canvrno-oai)\n\n- [#21089](https://github.com/openai/codex/pull/21089) [codex] Fix fork --last cwd filtering [@canvrno-oai](https://github.com/canvrno-oai)\n\n- [#21152](https://github.com/openai/codex/pull/21152) revert legacy notify deprecation [@abhinav-oai](https://github.com/abhinav-oai)\n\n- [#21190](https://github.com/openai/codex/pull/21190) fix(tui): external editor expansion for same-size large pastes [@fcoury-oai](https://github.com/fcoury-oai)\n\n- [#20111](https://github.com/openai/codex/pull/20111) fix(sandboxing): Bound advisory system bwrap startup probe [@viyatb-oai](https://github.com/viyatb-oai)\n\n- [#21220](https://github.com/openai/codex/pull/21220) chore: add minimal proxy egress diagnostics [@viyatb-oai](https://github.com/viyatb-oai)\n\n- [#20819](https://github.com/openai/codex/pull/20819) feat(tui): add raw scrollback mode [@fcoury-oai](https://github.com/fcoury-oai)\n\n- [#21225](https://github.com/openai/codex/pull/21225) app-server: ignore persist_extended_history param [@owenlin0](https://github.com/owenlin0)\n\n- [#17089](https://github.com/openai/codex/pull/17089) [codex-analytics] add tool item event schemas [@rhan-oai](https://github.com/rhan-oai)\n\n- [#20647](https://github.com/openai/codex/pull/20647) Route process tools to selected environments [@starr-openai](https://github.com/starr-openai)\n\n- [#20321](https://github.com/openai/codex/pull/20321) hook trust metadata and enforcement [@abhinav-oai](https://github.com/abhinav-oai)\n\n- [#21221](https://github.com/openai/codex/pull/21221) [codex] Use shared app-server JSON-RPC error helpers [@pakrym-oai](https://github.com/pakrym-oai)\n\n- [#21063](https://github.com/openai/codex/pull/21063) add turn items view to app-server turns [@rhan-oai](https://github.com/rhan-oai)\n\n- [#21001](https://github.com/openai/codex/pull/21001) feat(tui): route /diff through workspace commands [@fcoury-oai](https://github.com/fcoury-oai)\n\n- [#20065](https://github.com/openai/codex/pull/20065) feat(tui): redesign session picker [@fcoury-oai](https://github.com/fcoury-oai)\n\n- [#21127](https://github.com/openai/codex/pull/21127) fix(linux-sandbox): avoid panic on bwrap build failures [@viyatb-oai](https://github.com/viyatb-oai)\n\n- [#21234](https://github.com/openai/codex/pull/21234) fix(linux-sandbox): isolate Linux sandbox synthetic mount registry per user for shared codex use case [@viyatb-oai](https://github.com/viyatb-oai)\n\n- [#20687](https://github.com/openai/codex/pull/20687) [codex] Split tool handlers by tool name [@pakrym-oai](https://github.com/pakrym-oai)\n\n- [#21113](https://github.com/openai/codex/pull/21113) Auto-deny MCP elicitations for Xcode 26.4 clients [@etraut-openai](https://github.com/etraut-openai)\n\n- [#21243](https://github.com/openai/codex/pull/21243) [codex] fix TUI turn items view fixtures [@fcoury-oai](https://github.com/fcoury-oai)\n\n- [#21146](https://github.com/openai/codex/pull/21146) Enable V8 sandboxing for source-built builds [@cconger](https://github.com/cconger)\n\n- [#20689](https://github.com/openai/codex/pull/20689) Inject state DB, agent graph store [@rasmusrygaard](https://github.com/rasmusrygaard)\n\n- [#19575](https://github.com/openai/codex/pull/19575) Add cloud executor registration to exec-server [@miz-openai](https://github.com/miz-openai)\n\n- [#20577](https://github.com/openai/codex/pull/20577) codex: use ThreadStore history for core review forks [@wiltzius-openai](https://github.com/wiltzius-openai)\n\n- [#21261](https://github.com/openai/codex/pull/21261) fix build [@bolinfest](https://github.com/bolinfest)\n\n- [#21251](https://github.com/openai/codex/pull/21251) chore(app-server-protocol): split v2 API definitions into modules [@owenlin0](https://github.com/owenlin0)\n\n- [#21259](https://github.com/openai/codex/pull/21259) ci: trigger rusty-v8 releases from tags [@cconger](https://github.com/cconger)\n\n- [#21255](https://github.com/openai/codex/pull/21255) linux-sandbox: use standalone bundled bwrap [@bolinfest](https://github.com/bolinfest)\n\n- [#21256](https://github.com/openai/codex/pull/21256) release: publish standalone bwrap artifacts [@bolinfest](https://github.com/bolinfest)\n\n- [#21260](https://github.com/openai/codex/pull/21260) [codex] Move thread naming to app server [@pakrym-oai](https://github.com/pakrym-oai)\n\n- [#21219](https://github.com/openai/codex/pull/21219) Add model and reasoning effort to MCP turn metadata [@mchen-oai](https://github.com/mchen-oai)\n\n- [#21275](https://github.com/openai/codex/pull/21275) Share Git safe-command logic on Windows [@iceweasel-oai](https://github.com/iceweasel-oai)\n\n- [#21257](https://github.com/openai/codex/pull/21257) release/npm: bundle standalone bwrap on Linux [@bolinfest](https://github.com/bolinfest)\n\n- [#21276](https://github.com/openai/codex/pull/21276) [codex] Remove unused ListModels op [@pakrym-oai](https://github.com/pakrym-oai)\n\n- [#21282](https://github.com/openai/codex/pull/21282) [codex] Remove legacy ListSkills op [@pakrym-oai](https://github.com/pakrym-oai)\n\n- [#21271](https://github.com/openai/codex/pull/21271) Expose plugin manifest keywords in app server [@alfozan](https://github.com/alfozan)\n\n- [#20949](https://github.com/openai/codex/pull/20949) [codex-analytics] rework thread_source for thread analytics [@rhan-oai](https://github.com/rhan-oai)\n\n- [#21124](https://github.com/openai/codex/pull/21124) feat: Add plugin share access controls [@xl-openai](https://github.com/xl-openai)\n\n- [#20724](https://github.com/openai/codex/pull/20724) app-server: align dynamic tool identifiers with Responses API [@eternal-openai](https://github.com/eternal-openai)\n\n- [#21055](https://github.com/openai/codex/pull/21055) Preserve session MCP config on refresh [@aaronl-openai](https://github.com/aaronl-openai)\n\n- [#21277](https://github.com/openai/codex/pull/21277) [mcp] Return Accept early per feedback. [@mzeng-openai](https://github.com/mzeng-openai)\n\n- [#21285](https://github.com/openai/codex/pull/21285) fix(bwrap): emit libcap after standalone archive [@viyatb-oai](https://github.com/viyatb-oai)\n\n- [#21312](https://github.com/openai/codex/pull/21312) release: bundle bwrap with Linux codex DotSlash artifact [@bolinfest](https://github.com/bolinfest)\n\n- [#19193](https://github.com/openai/codex/pull/19193) Support Codex Apps auth elicitations [@mzeng-openai](https://github.com/mzeng-openai)\n\n- [#20437](https://github.com/openai/codex/pull/20437) feat: add `session_id` [@jif-oai](https://github.com/jif-oai)\n\n- [#21328](https://github.com/openai/codex/pull/21328) test: isolate app-server-client in-process test state [@jif-oai](https://github.com/jif-oai)\n\n- [#21329](https://github.com/openai/codex/pull/21329) feat: include thread ID in MCP turn metadata [@jif-oai](https://github.com/jif-oai)\n\n- [#21332](https://github.com/openai/codex/pull/21332) feat: return session ID from thread/fork [@jif-oai](https://github.com/jif-oai)\n\n- [#21337](https://github.com/openai/codex/pull/21337) Revert \"feat: support template interpolation in multi-agent usage hints\" [@jif-oai](https://github.com/jif-oai)\n\n- [#21249](https://github.com/openai/codex/pull/21249) Propagate cache key and service tiers in compact [@aibrahim-oai](https://github.com/aibrahim-oai)\n\n- [#21182](https://github.com/openai/codex/pull/21182) Move installation ID resolution out of core startup [@jif-oai](https://github.com/jif-oai)\n\n- [#21214](https://github.com/openai/codex/pull/21214) chore: spawn MCP for memories [@jif-oai](https://github.com/jif-oai)\n\n- [#21336](https://github.com/openai/codex/pull/21336) feat(app-server): move v2 `sessionId` onto `Thread` [@jif-oai](https://github.com/jif-oai)\n\n- [#21350](https://github.com/openai/codex/pull/21350) [codex] fix builtin MCP Windows path test [@jif-oai](https://github.com/jif-oai)\n\n- [#20971](https://github.com/openai/codex/pull/20971) 2- Use string service tiers in session protocol [@aibrahim-oai](https://github.com/aibrahim-oai)\n\n- [#21278](https://github.com/openai/codex/pull/21278) Move message history out of core [@pakrym-oai](https://github.com/pakrym-oai)\n\n- [#21284](https://github.com/openai/codex/pull/21284) [codex] Add response.processed websocket request [@pakrym-oai](https://github.com/pakrym-oai)\n\n- [#21367](https://github.com/openai/codex/pull/21367) rollout: coalesce thread updated_at touches [@jif-oai](https://github.com/jif-oai)\n\n- [#21378](https://github.com/openai/codex/pull/21378) feat: move auto vaccum [@jif-oai](https://github.com/jif-oai)\n\n- [#21263](https://github.com/openai/codex/pull/21263) [codex] Coordinate OpenAI docs sample with API key setup [@mifan-oai](https://github.com/mifan-oai)\n\n- [#21351](https://github.com/openai/codex/pull/21351) fix(tui): keep Ctrl-C stashed drafts after /clear [@fcoury-oai](https://github.com/fcoury-oai)\n\n- [#21389](https://github.com/openai/codex/pull/21389) vendor: update bubblewrap to 0.11.2 [@bolinfest](https://github.com/bolinfest)\n\n- [#21281](https://github.com/openai/codex/pull/21281) Remove core MCP list tools op [@pakrym-oai](https://github.com/pakrym-oai)\n\n- [#21381](https://github.com/openai/codex/pull/21381) [codex] Handle git pagination flags by position [@iceweasel-oai](https://github.com/iceweasel-oai)\n\n- [#21397](https://github.com/openai/codex/pull/21397) fix(tui): persist ctrl-c draft via app event [@fcoury-oai](https://github.com/fcoury-oai)\n\n- [#19431](https://github.com/openai/codex/pull/19431) Route opted-in MCP elicitations through Guardian [@cd-oai](https://github.com/cd-oai)\n\n- [#21107](https://github.com/openai/codex/pull/21107) Avoid noisy OTEL diagnostics in codex exec [@cpaasch-oai](https://github.com/cpaasch-oai)\n\n- [#21390](https://github.com/openai/codex/pull/21390) Avoid hard-coded environment context shell [@starr-openai](https://github.com/starr-openai)\n\n- [#21090](https://github.com/openai/codex/pull/21090) [codex] Dedupe fallback model metadata warnings [@canvrno-oai](https://github.com/canvrno-oai)\n\n- [#21395](https://github.com/openai/codex/pull/21395) [codex] Split tool handlers into separate files [@pakrym-oai](https://github.com/pakrym-oai)\n\n- [#21401](https://github.com/openai/codex/pull/21401) [codex-tui] pass thread source for tui threads [@rhan-oai](https://github.com/rhan-oai)\n\n- [#17090](https://github.com/openai/codex/pull/17090) [codex-analytics] emit tool item events from item lifecycle [@rhan-oai](https://github.com/rhan-oai)\n\n- [#21409](https://github.com/openai/codex/pull/21409) [codex] Fix Windows sandbox git safe.directory for worktrees [@iceweasel-oai](https://github.com/iceweasel-oai)\n\n- [#21379](https://github.com/openai/codex/pull/21379) Document Codex git commit attribution config [@henzelmann-oai](https://github.com/henzelmann-oai)\n\n- [#21287](https://github.com/openai/codex/pull/21287) Move skills watcher to app-server [@pakrym-oai](https://github.com/pakrym-oai)\n\n- [#21416](https://github.com/openai/codex/pull/21416) [codex] Move tool specs into core handlers [@pakrym-oai](https://github.com/pakrym-oai)\n\n- [#21419](https://github.com/openai/codex/pull/21419) feat: Add marketplace source filtering and plugin share context [@xl-openai](https://github.com/xl-openai)\n\n- [#19905](https://github.com/openai/codex/pull/19905) Add compact lifecycle hooks (started by vincentkoc - external contrib) [@eternal-openai](https://github.com/eternal-openai)\n\n- [#21460](https://github.com/openai/codex/pull/21460) Revert \"Move skills watcher to app-server\" [@pakrym-oai](https://github.com/pakrym-oai)\n\n- [#21450](https://github.com/openai/codex/pull/21450) fix(tui): clear first inline viewport render [@fcoury-oai](https://github.com/fcoury-oai)\n\n- [#21427](https://github.com/openai/codex/pull/21427) [codex] Delete tool handler plan indirection [@pakrym-oai](https://github.com/pakrym-oai)\n\n- [#21423](https://github.com/openai/codex/pull/21423) [codex] Add OpenAI Developers to tool suggest allowlist [@mifan-oai](https://github.com/mifan-oai)\n\n- [#21340](https://github.com/openai/codex/pull/21340) [codex] allow shared config reads in app-server queue [@xli-oai](https://github.com/xli-oai)\n\n- [#21441](https://github.com/openai/codex/pull/21441) [codex] Parallelize skills list cwd loading [@xli-oai](https://github.com/xli-oai)\n\n- [#21481](https://github.com/openai/codex/pull/21481) Revert state DB injection and agent graph store [@pakrym-oai](https://github.com/pakrym-oai)","publishedAt":"2026-05-07T00:00:00.000Z","url":"https://developers.openai.com/codex/changelog/#github-release-319066904","media":[],"prerelease":false,"source":{"slug":"openai-codex-changelog","name":"Codex Changelog","type":"feed"},"org":{"slug":"openai","name":"OpenAI"},"product":{"slug":"codex","name":"Codex"}},{"id":"rel_hDzvy12saDKSq_Y9Ws6DV","version":"2.1.132","type":"feature","title":"2.1.132","summary":"- Added `CLAUDE_CODE_SESSION_ID` environment variable to the Bash tool subprocess environment, matching the `session_id` passed to hooks\n\n- Added `CLA...","content":"- Added `CLAUDE_CODE_SESSION_ID` environment variable to the Bash tool subprocess environment, matching the `session_id` passed to hooks\n\n- Added `CLAUDE_CODE_DISABLE_ALTERNATE_SCREEN=1` env var to opt out of the fullscreen alternate-screen renderer and keep the conversation in the terminal's native scrollback\n\n- Added a \"Pasting…\" footer hint while a Ctrl+V image paste is being read from the clipboard\n\n- Fixed external SIGINT (e.g. IDE stop button, `kill -INT`) not running graceful shutdown — terminal modes are now restored and the `--resume` hint is printed instead of an abrupt exit\n\n- Fixed an uncaught exception when the terminal is closed or SSH disconnects mid-session under the native build\n\n- Fixed `--resume` failing with `no low surrogate in string` when a tool error truncation split an emoji; pre-corrupted sessions are sanitized on load\n\n- Fixed `--permission-mode` flag being ignored when resuming a plan-mode session with `-p --continue`/`--resume`, and plan mode not being re-applied after `ExitPlanMode` within the same session\n\n- Fixed fullscreen mode showing a blank screen after laptop sleep/wake or Ctrl+Z/`fg` until the next keystroke or stream output\n\n- Fixed cursor landing mid-grapheme on Ctrl+E/A/K/U/arrow keys when an Indic conjunct or ZWJ emoji wraps across lines\n\n- Fixed vim operators corrupting text containing decomposed (NFD) accented characters\n\n- Fixed pasting text starting with `/` silently swallowing the input or triggering an unknown-command reply\n\n- Fixed pasting dumping stray escape sequences into the prompt when focus events or mouse-tracking reports interleave with the bracketed paste\n\n- Fixed mouse wheel scrolling being too fast in Cursor and VS Code 1.92–1.104 due to an upstream xterm.js bug\n\n- Fixed scroll-wheel handling in JetBrains IDE 2025.2 terminals (spurious arrow keys, wrong-direction events, runaway acceleration)\n\n- Fixed `/usage` Ctrl+S hanging when copying the stats screenshot to the clipboard on Linux/X11\n\n- Fixed `/terminal-setup` showing a contradictory error in Windows Terminal — Shift+Enter is natively supported there\n\n- Fixed `/effort` picker not reflecting the `CLAUDE_CODE_EFFORT_LEVEL` env var override\n\n- Fixed `/status` showing the wrong default model for some users\n\n- Fixed slash command autocomplete popup being capped at ~3–5 visible commands instead of scaling with terminal height\n\n- Fixed statusline `context_window` token counts reflecting cumulative session totals instead of current context usage\n\n- Fixed Alt+T (thinking toggle) not working on macOS terminals without \"Option as Meta\" enabled (iTerm2, Terminal.app defaults)\n\n- Fixed dead keyboard input on Windows after re-opening a background session from `claude agents`\n\n- Fixed unbounded memory growth (10GB+ RSS) when a stdio MCP server writes non-protocol data to stdout\n\n- Fixed MCP servers that connect but fail `tools/list` silently showing 0 tools — they now retry once and show \"connected · tools fetch failed\" in `/mcp`\n\n- Fixed unauthorized claude.ai MCP connectors showing as \"failed\" instead of \"needs auth\", and headless `-p` mode retrying non-transient 4xx connection failures\n\n- Improved visual consistency in slash command dialogs and `/login`, `/upgrade`, `/extra-usage` dialog spacing\n\n- Updated the `/tui fullscreen` startup banner to describe additional renderer benefits (lower memory usage, mouse support, auto-copy on select)\n\n- Fixed Bedrock and Vertex 400 errors when `ENABLE_PROMPT_CACHING_1H` is set","publishedAt":"2026-05-06T22:28:20.000Z","url":"https://code.claude.com/docs/en/changelog#2-1-132","media":[],"prerelease":false,"source":{"slug":"claude-code","name":"Claude Code","type":"scrape"},"org":{"slug":"anthropic","name":"Anthropic"},"product":{"slug":"claude-code","name":"Claude Code"}},{"id":"rel_j0WgpgR4EIG50wKVrUVrG","version":"v2.35.1","type":"feature","title":"v2.35.1","summary":"## 2.35.1 (2026-05-06)\n\nFull Changelog: [v2.35.0...v2.35.1](https://github.com/openai/openai-python/compare/v2.35.0...v2.35.1)\n\n### Bug Fixes\n\n* **api...","content":"## 2.35.1 (2026-05-06)\n\nFull Changelog: [v2.35.0...v2.35.1](https://github.com/openai/openai-python/compare/v2.35.0...v2.35.1)\n\n### Bug Fixes\n\n* **api:** fix imagegen `size` enum regression ([4484653](https://github.com/openai/openai-python/commit/44846536bc3b02c393daa5bae70a85de04c7f621))","publishedAt":"2026-05-06T21:37:58.000Z","url":"https://github.com/openai/openai-python/releases/tag/v2.35.1","media":[],"prerelease":false,"source":{"slug":"openai-python-sdk","name":"Python SDK","type":"github"},"org":{"slug":"openai","name":"OpenAI"},"product":null},{"id":"rel_4bi306cfVD5TcVoR2olFj","version":"v2.35.0","type":"feature","title":"v2.35.0","summary":"## 2.35.0 (2026-05-06)\n\nFull Changelog: [v2.34.0...v2.35.0](https://github.com/openai/openai-python/compare/v2.34.0...v2.35.0)\n\n### Features\n\n* **api:...","content":"## 2.35.0 (2026-05-06)\n\nFull Changelog: [v2.34.0...v2.35.0](https://github.com/openai/openai-python/compare/v2.34.0...v2.35.0)\n\n### Features\n\n* **api:** update image 2 ([0ba55d7](https://github.com/openai/openai-python/commit/0ba55d7569565045426e1587906a70d5682a4bba))\n* **api:** manual updates ([72bf67a](https://github.com/openai/openai-python/commit/72bf67acbc9f030c20db3d5a1a74ea6d67d55f51))\n\n\n### Chores\n\n* remove legacy python cli ([32f36e4](https://github.com/openai/openai-python/commit/32f36e447d02c3124af8ab48fcc3537df2fed66e))\n* rename legacy python cli entrypoint ([a3b182d](https://github.com/openai/openai-python/commit/a3b182d6d2c2e6fe1d53ca7550b2d43e0f8b2cd3))\n\n\n### Documentation\n\n* **api:** update top_logprobs parameter description across chat and responses ([f9d339f](https://github.com/openai/openai-python/commit/f9d339fcea63feaa1bdf918a4599f2b032c83517))","publishedAt":"2026-05-06T16:36:40.000Z","url":"https://github.com/openai/openai-python/releases/tag/v2.35.0","media":[],"prerelease":false,"source":{"slug":"openai-python-sdk","name":"Python SDK","type":"github"},"org":{"slug":"openai","name":"OpenAI"},"product":null},{"id":"rel_g38BQhUrDSmX7kEcsj10n","version":"2.1.131","type":"feature","title":"2.1.131","summary":"- Fixed VS Code extension failing to activate on Windows due to a hardcoded build path in the bundled SDK (`createRequire` polyfill bug)\n\n- Fixed Mant...","content":"- Fixed VS Code extension failing to activate on Windows due to a hardcoded build path in the bundled SDK (`createRequire` polyfill bug)\n\n- Fixed Mantle endpoint authentication failing with missing `x-api-key` header","publishedAt":"2026-05-06T07:52:26.000Z","url":"https://code.claude.com/docs/en/changelog#2-1-131","media":[],"prerelease":false,"source":{"slug":"claude-code","name":"Claude Code","type":"scrape"},"org":{"slug":"anthropic","name":"Anthropic"},"product":{"slug":"claude-code","name":"Claude Code"}},{"id":"rel_LsngR0i8jVRYgYJi0JBWa","version":"2.1.129","type":"feature","title":"2.1.129","summary":"- Added `--plugin-url ` flag to fetch a plugin `.zip` archive from a URL for the current session\n\n- Added `CLAUDE_CODE_FORCE_SYNC_OUTPUT=1` env var to...","content":"- Added `--plugin-url ` flag to fetch a plugin `.zip` archive from a URL for the current session\n\n- Added `CLAUDE_CODE_FORCE_SYNC_OUTPUT=1` env var to force-enable synchronized output on terminals that auto-detection misses (e.g. Emacs `eat`)\n\n- Added `CLAUDE_CODE_PACKAGE_MANAGER_AUTO_UPDATE`: when set on Homebrew or WinGet installations, Claude Code runs the upgrade command in the background and prompts to restart\n\n- Plugin manifests: `themes` and `monitors` should now be declared under `\"experimental\": { ... }`. Top-level declarations still work but `claude plugin validate` will warn\n\n- Gateway `/v1/models` discovery for the `/model` picker is now opt-in via `CLAUDE_CODE_ENABLE_GATEWAY_MODEL_DISCOVERY=1` (was automatic in 2.1.126–2.1.128)\n\n- Ctrl+R history picker now defaults to searching all prompts across all projects, matching pre-2.1.124 behavior. Press Ctrl+S to narrow to the current project or session\n\n- Third-party deployments (Bedrock, Vertex, Foundry, or `ANTHROPIC_BASE_URL` gateway) no longer see spinner tips pointing at first-party Anthropic surfaces\n\n- `skillOverrides` setting now works: `off` hides from model and `/`, `user-invocable-only` hides from model only, `name-only` collapses description\n\n- The `claude_code.pull_request.count` OTel metric now counts PRs/MRs created via MCP tools, not just shell commands\n\n- Policy refusal error messages now include the API Request ID for easier support debugging\n\n- Fixed API errors with unrecognized 400 status codes showing raw JSON instead of the underlying error message\n\n- Fixed `/clear` not resetting the terminal tab title after a conversation\n\n- Fixed session title chip from `/rename` disappearing while a permission or other dialog is active\n\n- Fixed agent panel below the prompt being hidden when subagents are running (regression in 2.1.122)\n\n- Fixed external-editor handoff (Ctrl+G) blanking the conversation history above the prompt\n\n- Fixed `/context` dumping its rendered ASCII visualization grid into the conversation, wasting ~1.6k tokens per call\n\n- Fixed `/agents` Library list arrow-key navigation: the highlighted agent now stays visible when the list exceeds the viewport\n\n- Fixed `/branch` success message not including the new branch's session id for `/resume`\n\n- Fixed bold headers with keycap/ZWJ/skin-tone emoji losing trailing characters in fullscreen mode\n\n- Fixed server-managed settings policy not applying for enterprise/team users whose stored OAuth credentials lacked the `user:inference` scope\n\n- Fixed OAuth refresh race after wake-from-sleep that could log out all running sessions\n\n- Fixed 1-hour prompt cache TTL being silently downgraded to 5 minutes\n\n- Fixed cache-miss warning appearing spuriously after `/clear` or compaction when changing `/effort` or `/model`\n\n- Fixed `Bash(mkdir *)`, `Bash(touch *)` and similar allow rules not being honored for in-project paths\n\n- Fixed `deniedMcpServers` patterns with a `*://` scheme wildcard not matching mixed-case hostnames\n\n- Fixed harmless WebSocket warning being logged as an error in `--debug` during voice mode\n\n- [VSCode] Fixed `/clear` not clearing the conversation context and displayed transcript","publishedAt":"2026-05-06T01:45:28.000Z","url":"https://code.claude.com/docs/en/changelog#2-1-129","media":[],"prerelease":false,"source":{"slug":"claude-code","name":"Claude Code","type":"scrape"},"org":{"slug":"anthropic","name":"Anthropic"},"product":{"slug":"claude-code","name":"Claude Code"}},{"id":"rel_aK8OHrR6SUuietqgaYEn3","version":null,"type":"feature","title":"Codex analytics governance docs update","summary":"# Codex analytics governance docs update\n\nUpdated the Codex enterprise governance guide with more detailed coverage of the\nAnalytics dashboard charts,...","content":"# Codex analytics governance docs update\n\nUpdated the Codex enterprise governance guide with more detailed coverage of the\nAnalytics dashboard charts, data export options, and enterprise Analytics API\nendpoints.","publishedAt":"2026-05-06T00:00:00.000Z","url":"https://developers.openai.com/codex/changelog/#codex-2026-05-06","media":[],"prerelease":false,"source":{"slug":"openai-codex-changelog","name":"Codex Changelog","type":"feed"},"org":{"slug":"openai","name":"OpenAI"},"product":{"slug":"codex","name":"Codex"}},{"id":"rel_Lb0HlwkVrFWMibnOrpFqa","version":null,"type":"feature","title":"Multiagent sessions and Outcomes launched in public beta","summary":"Multiagent sessions and Outcomes are now in public beta under the standard `managed-agents-2026-04-01` beta header. Vault credential background refres...","content":"Multiagent sessions and Outcomes are now in public beta under the standard `managed-agents-2026-04-01` beta header. Vault credential background refresh is now supported for `mcp_oauth` credentials. Webhooks for Claude Managed Agents are now supported, with event types including session and vault lifecycle events. Additional filtering and sorting options are now supported—sessions can be filtered by status, and events can be filtered by type and creation time.","publishedAt":"2026-05-06T00:00:00.000Z","url":"https://platform.claude.com/docs/en/release-notes/overview#multiagent-sessions-and-outcomes-launched-in-public-beta","media":[],"prerelease":false,"source":{"slug":"claude-platform","name":"Claude Platform","type":"scrape"},"org":{"slug":"anthropic","name":"Anthropic"},"product":{"slug":"claude-platform","name":"Claude Platform"}},{"id":"rel_wazXJ5EwI6rLsw2E-q-GW","version":"5.5","type":"feature","title":"GPT-5.5 Instant: smarter, clearer, and more personalized","summary":"GPT-5.5 Instant updates ChatGPT’s default model with smarter, more accurate answers, reduced hallucinations, and improved personalization controls.","content":"GPT-5.5 Instant updates ChatGPT’s default model with smarter, more accurate answers, reduced hallucinations, and improved personalization controls.","publishedAt":"2026-05-05T10:00:00.000Z","url":"https://openai.com/index/gpt-5-5-instant","media":[],"prerelease":false,"source":{"slug":"openai-news","name":"Product Releases","type":"feed"},"org":{"slug":"openai","name":"OpenAI"},"product":null},{"id":"rel_ll5MnZFK1GFwP1O8U2e0j","version":null,"type":"feature","title":"Realtime prompting guide","summary":"Shows how to prompt realtime voice agents, including Realtime 2 reasoning, preambles, tools, and exact entity capture.","content":"Shows how to prompt realtime voice agents, including Realtime 2 reasoning, preambles, tools, and exact entity capture.","publishedAt":"2026-05-05T00:00:00.000Z","url":"https://platform.openai.com/docs/guides/realtime-models-prompting","media":[],"prerelease":false,"source":{"slug":"openai-api-changelog","name":"API Changelog","type":"feed"},"org":{"slug":"openai","name":"OpenAI"},"product":{"slug":"api","name":"API"}},{"id":"rel_djznJQwUdyNAxK8BMQvss","version":null,"type":"feature","title":"New ways to buy ChatGPT ads","summary":"OpenAI expands ChatGPT ads with a beta self-serve Ads Manager, CPC bidding, and enhanced measurement tools—built to protect privacy and keep conversat...","content":"OpenAI expands ChatGPT ads with a beta self-serve Ads Manager, CPC bidding, and enhanced measurement tools—built to protect privacy and keep conversations separate from ads.","publishedAt":"2026-05-05T00:00:00.000Z","url":"https://openai.com/index/new-ways-to-buy-chatgpt-ads","media":[],"prerelease":false,"source":{"slug":"openai-news","name":"Product Releases","type":"feed"},"org":{"slug":"openai","name":"OpenAI"},"product":null}],"pagination":{"nextCursor":"2026-05-05T00:00:00.000Z|2026-05-05T22:01:10.293Z|rel_djznJQwUdyNAxK8BMQvss","limit":20}}