{"slug":"coding-agents","name":"Coding Agents","description":"AI pair programmers and autonomous coding agents — IDE-integrated assistants and cloud agents that read, edit, and ship code.","isFeatured":true,"members":[{"kind":"product","slug":"claude-code","name":"Claude Code","description":"AI coding agent for the terminal and IDE","org":{"slug":"anthropic","name":"Anthropic","domain":"anthropic.com","avatarUrl":"https://media.releases.sh/orgs/anthropic.png","githubHandle":"anthropics"}},{"kind":"product","slug":"codex","name":"Codex","description":null,"org":{"slug":"openai","name":"OpenAI","domain":"openai.com","avatarUrl":"https://media.releases.sh/orgs/openai.png","githubHandle":"openai"}},{"kind":"org","slug":"cursor","name":"Cursor","domain":"cursor.com","avatarUrl":"https://media.releases.sh/orgs/cursor.png","githubHandle":"getcursor","description":"AI-powered code editor built on VS Code that integrates large language models directly into the development workflow."},{"kind":"product","slug":"github-copilot","name":"GitHub Copilot","description":null,"org":{"slug":"github","name":"GitHub","domain":"github.com","avatarUrl":"https://media.releases.sh/orgs/github.png","githubHandle":"github"}},{"kind":"product","slug":"devin","name":"Devin","description":null,"org":{"slug":"cognition","name":"Cognition","domain":"cognition.ai","avatarUrl":"https://media.releases.sh/orgs/cognition.png","githubHandle":"Cognition-ai"}},{"kind":"org","slug":"windsurf","name":"Windsurf","domain":"windsurf.com","avatarUrl":"https://media.releases.sh/orgs/windsurf.png","githubHandle":"Exafunction","description":"AI-powered code editor that keeps developers in flow state with local and cloud coding agents."}],"orgs":[{"slug":"cursor","name":"Cursor","domain":"cursor.com","avatarUrl":"https://media.releases.sh/orgs/cursor.png","githubHandle":"getcursor","description":"AI-powered code editor built on VS Code that integrates large language models directly into the development workflow."},{"slug":"windsurf","name":"Windsurf","domain":"windsurf.com","avatarUrl":"https://media.releases.sh/orgs/windsurf.png","githubHandle":"Exafunction","description":"AI-powered code editor that keeps developers in flow state with local and cloud coding agents."}],"releases":[{"id":"rel_OlVXj5Cj47FgjHJZQl7zn","version":"v2.1.148","type":"feature","title":"v2.1.148","summary":"Fixed a regression introduced in 2.1.147 where the Bash tool was returning exit code 127 on every command for some users.","titleGenerated":"Claude Code v2.1.148 fixes Bash tool exit code 127 regression","titleShort":"Bash tool exit code 127 regression fixed","content":"## What's changed\n\n- Fixed the Bash tool returning exit code 127 on every command for some users (a regression introduced in 2.1.147)\n","publishedAt":"2026-05-22T01:16:52.000Z","url":"https://code.claude.com/docs/en/changelog#2-1-148","media":[],"prerelease":false,"source":{"slug":"claude-code","name":"Claude Code","type":"scrape"},"org":{"slug":"anthropic","name":"Anthropic"},"product":{"slug":"claude-code","name":"Claude Code"},"coverageCount":0},{"id":"rel_taDieQOnTX4-gXazi5aA5","version":"v2.1.147","type":"feature","title":"v2.1.147","summary":"Added the Workflow tool for deterministic multi-agent orchestration, disabled by default via CLAUDE_CODE_WORKFLOWS=1. Hardened REPL and Workflow tool sandboxes against prototype-pollution and thenable-based escapes. Renamed /simplify to /code-review with configurable effort levels and GitHub PR comment support; the old cleanup-and-fix behavior is removed. Pinned background sessions now persist when idle and restart in-place for updates. Plus dozens of smaller fixes across shell snapshots, PowerShell tool output, Windows Terminal rendering, and plugin management.","titleGenerated":"Claude Code v2.1.147 adds Workflow tool and hardens sandbox security","titleShort":"Workflow tool for multi-agent orchestration; sandbox escapes hardened","content":"## What's changed\n\n- Added the `Workflow` tool for deterministic multi-agent orchestration. It is off by default — set `CLAUDE_CODE_WORKFLOWS=1` to enable\n- Pinned background sessions (`Ctrl+T` in `claude agents`) now stay alive when idle, are restarted in place to apply Claude Code updates, and are shed under memory pressure only after non-pinned sessions\n- Renamed `/simplify` to `/code-review`. It now reports correctness bugs at a chosen effort level (e.g., `/code-review high`); pass `--comment` to post findings as inline GitHub PR comments. The old cleanup-and-fix behavior has been removed\n- Hardened REPL and Workflow tool sandboxes against prototype-pollution and thenable-based escapes\n- Improved auto-updater: retries transient network failures, reports specific error categories and OS error codes on failure, and shows the current version when an update fails\n- Improved diff rendering performance for large file edits\n- Prompt history no longer records consecutive duplicate entries — recalling a prompt with arrow-up and submitting it again won't add another copy\n- Fixed enterprise login restrictions (`forceLoginOrgUUID` and `forceLoginMethod` managed-settings) not being enforced against third-party-provider and API-key sessions\n- Fixed `&` in `!` command output displaying as `&`, which broke copy-pasting URLs from commands like `gcloud auth login` on headless machines\n- Fixed unknown slash commands silently doing nothing in headless/SDK mode — they now show an error message\n- Fixed `/help` rendering a broken tab header and showing only one command per page on small terminals when not in fullscreen mode\n- Fixed shell snapshot dropping user functions whose names start with a single underscore, which broke aliases referencing them\n- Fixed plugin agents that declare multiple `Agent(...)` types in `tools:` frontmatter dropping all but the last entry\n- Fixed hook `if` conditions like `PowerShell(git push*)` never matching — only `PowerShell(*)` worked\n- Fixed PowerShell tool dropping output for commands that rely on the default formatter\n- Fixed: on Windows, \"Yes, and don't ask again\" for a PowerShell script invocation now writes a rule that actually matches on subsequent runs\n- Fixed PowerShell tool failing on Windows with exit code 1 when `pwsh` is installed via winget or the Microsoft Store\n- Fixed `/effort` opening with the slider on the wrong level — it now starts at your current effort\n- Fixed paginating MCP servers dropping resources, templates, and prompts past page 1\n- Fixed full-screen strobing in attached background sessions on Windows Terminal while Claude is streaming\n- Fixed: on Windows, removing a background-job worktree no longer follows NTFS junctions into the main repo\n- Fixed `/background` refusing sessions whose only typed input was a skill or custom slash command\n- Fixed auto mode suppressing `AskUserQuestion` when the user or a skill explicitly relies on it; the auto-mode classifier now sees the user's answers as intent signal\n- Fixed `/theme` \"New custom theme\" and color editor dialogs not responding to Esc\n- Fixed an uncaught exception at the end of streaming sessions when running via the Agent SDK\n- Fixed a rare hang when waiting for scroll to settle on Windows\n- Fixed stale and doubled rows in the agent view list on Windows when background session results contain wide (CJK) characters\n- Fixed pasted text being delivered to agents as an unreadable `[Pasted text #N]` placeholder instead of the actual content\n- Fixed plugin component counts in `claude plugin details` and `/plugin` being doubled when a plugin's manifest listed paths overlapping its default directories\n- Fixed backgrounded sessions re-prompting for tool permissions you already granted with \"don't ask again\"\n- Fixed GNOME Terminal right-click and middle-click paste not inserting text\n- Fixed `CLAUDE_CODE_SUBAGENT_MODEL` not applying to teammate processes spawned by agent teams\n- Fixed slash commands followed by a tab or newline being treated as an unknown command\n- Fixed several spacing and layout glitches in the `/plugin`, `/status`, `/mobile`, `/sandbox`, and `/permissions` menus\n- Fixed stripped images prompting the model to repeatedly re-read media that was no longer present\n","publishedAt":"2026-05-21T20:39:22.000Z","url":"https://code.claude.com/docs/en/changelog#2-1-147","media":[],"prerelease":false,"source":{"slug":"claude-code","name":"Claude Code","type":"scrape"},"org":{"slug":"anthropic","name":"Anthropic"},"product":{"slug":"claude-code","name":"Claude Code"},"coverageCount":0},{"id":"rel_7xjEpeW_oIjF6o1lESKp0","version":"v2.1.146","type":"feature","title":"v2.1.146","summary":"Fixed a Windows PowerShell tool regression where commands would fail with \"command line is invalid\" when pwsh is installed via winget or the Microsoft Store, and MCP resources/list, resources/templates/list, and prompts/list endpoints dropping items past page 1 on paginating servers. Also renamed /simplify to /code-review with optional effort levels, fixed full-screen strobing in attached background sessions on Windows Terminal, and improved auto-updater reliability with transient network failure retries.","titleGenerated":"Claude Code v2.1.146 fixes PowerShell tool and MCP pagination issues","titleShort":"PowerShell tool works with winget/Store; MCP pagination fixed","content":"## What's changed\n\n- Renamed `/simplify` to `/code-review` with an optional effort level (e.g. `/code-review high`)\n- Auto mode no longer suppresses `AskUserQuestion` when the user or a skill explicitly relies on it\n- Fixed Windows PowerShell tool failing with \"command line is invalid\" when `pwsh` is installed via winget or the Microsoft Store (regression in v2.1.124)\n- Fixed MCP `resources/list`, `resources/templates/list`, and `prompts/list` dropping items past page 1 on paginating servers\n- Fixed full-screen strobing in attached background sessions on Windows Terminal while Claude is streaming\n- Fixed the auto-updater status line not showing your current version when an update fails\n- Fixed on Windows, removing a background-job worktree no longer follows NTFS junctions into the main repo\n- Fixed `/background` refusing sessions whose only typed input was a skill or custom slash command\n- Fixed backgrounded sessions re-prompting for tool permissions you already granted with \"don't ask again\"\n- Fixed `/theme` color editor and \"New custom theme\" dialogs not responding to Esc\n- Fixed an uncaught exception at the end of streaming sessions when running via the Agent SDK\n- Fixed `forceLoginOrgUUID` and `forceLoginMethod` managed-settings policies not being enforced against third-party-provider and API-key sessions\n- Fixed GNOME Terminal right-click and middle-click paste not inserting text\n- Fixed `CLAUDE_CODE_SUBAGENT_MODEL` not being forwarded to child processes in multi-agent sessions\n- Improved auto-updater reliability: native version checks and downloads now retry transient network failures instead of failing immediately\n- Improved diff rendering performance for large file edits\n","publishedAt":"2026-05-21T01:51:52.000Z","url":"https://code.claude.com/docs/en/changelog#2-1-146","media":[],"prerelease":false,"source":{"slug":"claude-code","name":"Claude Code","type":"scrape"},"org":{"slug":"anthropic","name":"Anthropic"},"product":{"slug":"claude-code","name":"Claude Code"},"coverageCount":0},{"id":"rel_oTB6ALg9k686yOLzrEJYh","version":null,"type":"feature","title":"Appshots, goal mode, and more","summary":"Appshots are now available on macOS, allowing you to send the frontmost app window to Codex with a screenshot and available text by pressing both Command keys. Goal mode is no longer experimental and is available in the Codex app, IDE extension, and CLI. Remote computer use is now available, enabling Codex to use desktop apps after your Mac locks, including via Codex Mobile with safeguards such as short-lived authorization and covered displays. Also added plugin sharing through marketplace sources for ChatGPT Business, advanced in-app browser annotations for styling feedback, and browser-use improvements including faster image asset downloading, more effective structured data extraction, and improved reliability across Windows and other regions.","titleGenerated":"Codex adds appshots, promotes goal mode to general availability, and enables remote computer use","titleShort":"Appshots launch; goal mode exits beta; remote computer use enabled","content":"\\# Appshots, goal mode, and more \\[Appshots\\](/codex/appshots) are now available in the Codex app on macOS. Press both Command keys to send the frontmost app window to Codex with a screenshot and available text, so Codex can work from context in another app without you copying, pasting, or describing it manually. This launch also includes: - \\[Goal mode\\](/codex/prompting#goal-mode) is no longer an experimental feature and is available in the Codex app, IDE extension, and CLI. With Goal mode, you can have Codex drive toward a specific objective for hours or even days. - \\[Remote computer use\\](/codex/app/computer-use#locked-use), so Codex can use desktop apps after your Mac locks, including remotely via Codex Mobile. Codex scopes locked use to active, trusted computer use turns and includes safeguards such as short-lived authorization, covered displays, relock on local input, and manual-unlock fallback. - \\[Plugin sharing\\](/codex/plugins/build#share-a-local-plugin-with-your-workspace) through marketplace sources is available for ChatGPT Business. Enterprise support is coming soon. Teams can distribute reusable plugin bundles that include skills, app integrations, and MCP servers. - \\[Advanced in-app browser annotations\\](/codex/app/browser#styling-feedback) let you tweak styling such as font size, colors, and spacing directly using annotations. This gives Codex a clearer signal for changes. - Browser-use improvements across in-app browser & Chrome: - Codex can now download and extract all image assets from a page much more quickly. - Codex can now extract structured data from pages more effectively and find information more quickly with a read-only JS sandbox. - Chrome extension will create less clutter when using it. Codex will no longer create tab groups when taking over existing tabs, and at the end of a task for handoff. Instead, it uses tab icons to indicate status. - Significantly improved reliability for browser use. We fixed bugs on Windows, flaky availability of the plugin to non geo-blocked regions, and many other issues impacting performance.","publishedAt":"2026-05-21T00:00:00.000Z","url":"https://developers.openai.com/codex/changelog/#codex-2026-05-21","media":[],"prerelease":false,"source":{"slug":"openai-codex-changelog","name":"Codex Changelog","type":"feed"},"org":{"slug":"openai","name":"OpenAI"},"product":{"slug":"codex","name":"Codex"},"coverageCount":0},{"id":"rel_DM5h88vPQS4yA-C0K7_D5","version":"0.133.0","type":"feature","title":"Codex CLI Release: 0.133.0","summary":"Goals are now enabled by default, backed by dedicated storage, and track progress across active turns. Plugin discovery is easier to inspect with marketplace-aware list output, installed versions, and visible marketplace roots. Permission profiles gained list APIs, inheritance, managed requirements.toml support, and runtime refresh behavior. Fixed TUI startup choosing wrong working directory, plan-mode unexpected submit on modified Enter keys, stale background terminal poll events, and app-server startup/shutdown races.","titleGenerated":"Codex CLI v0.133.0 enables goals by default and extends plugin discovery","titleShort":"Goals now on by default; plugin discovery and permission profiles expanded","content":"## New Features\n\n- Goals are now enabled by default, backed by dedicated storage, and track progress across active turns. ([#23300](https://github.com/openai/codex/pull/23300), [#23685](https://github.com/openai/codex/pull/23685), [#23696](https://github.com/openai/codex/pull/23696), [#23732](https://github.com/openai/codex/pull/23732))\n- `codex remote-control` now runs like a foreground command, waits for readiness, reports machine status, and keeps explicit daemon-style `start`/`stop` commands. ([#22878](https://github.com/openai/codex/pull/22878))\n- Permission profiles gained list APIs, inheritance, managed `requirements.toml` support, runtime refresh behavior, and stronger Windows sandbox integration. ([#22928](https://github.com/openai/codex/pull/22928), [#23412](https://github.com/openai/codex/pull/23412), [#22270](https://github.com/openai/codex/pull/22270), [#23433](https://github.com/openai/codex/pull/23433), [#22931](https://github.com/openai/codex/pull/22931), [#23715](https://github.com/openai/codex/pull/23715))\n- Plugin discovery is easier to inspect, with marketplace-aware list output, installed versions, visible marketplace roots, and remote collection support. ([#23372](https://github.com/openai/codex/pull/23372), [#23584](https://github.com/openai/codex/pull/23584), [#23727](https://github.com/openai/codex/pull/23727), [#23730](https://github.com/openai/codex/pull/23730))\n- Extensions can observe more lifecycle events, including subagent start/stop, tool execution, turn metadata, and async approval/turn processing. ([#22782](https://github.com/openai/codex/pull/22782), [#22873](https://github.com/openai/codex/pull/22873), [#23309](https://github.com/openai/codex/pull/23309), [#23688](https://github.com/openai/codex/pull/23688), [#23690](https://github.com/openai/codex/pull/23690), [#23692](https://github.com/openai/codex/pull/23692))\n\n## Bug Fixes\n\n- Fixed TUI startup choosing the wrong working directory when reusing a local app-server socket. ([#23538](https://github.com/openai/codex/pull/23538))\n- Fixed plan-mode free-form answers so modified Enter keys, like Shift+Enter, no longer submit unexpectedly. ([#23536](https://github.com/openai/codex/pull/23536))\n- Removed stale background terminal poll events after a process exits. ([#23231](https://github.com/openai/codex/pull/23231))\n- Preserved raw code-mode exec output unless an explicit output token limit is requested. ([#23564](https://github.com/openai/codex/pull/23564))\n- Made AGENTS instruction loading more reliable, including local global reads and warnings for invalid UTF-8 instead of silent drops. ([#23343](https://github.com/openai/codex/pull/23343), [#23232](https://github.com/openai/codex/pull/23232))\n- Fixed app-server startup/shutdown races, empty resume/fork paths, plugin upgrade failures, and realtime v1 websocket compatibility. ([#23516](https://github.com/openai/codex/pull/23516), [#23578](https://github.com/openai/codex/pull/23578), [#23400](https://github.com/openai/codex/pull/23400), [#23356](https://github.com/openai/codex/pull/23356), [#23771](https://github.com/openai/codex/pull/23771))\n\n## Documentation\n\n- Added clearer plugin-creator guidance for updating and reinstalling local personal plugins. ([#23542](https://github.com/openai/codex/pull/23542))\n- Expanded app-server/API docs and schema coverage around managed permission profile requirements. ([#23433](https://github.com/openai/codex/pull/23433), [#23555](https://github.com/openai/codex/pull/23555))\n\n## Chores\n\n- Added a canonical Codex package archive pipeline and moved installers, npm packages, DotSlash, and SDK runtimes toward that shared layout. ([#23513](https://github.com/openai/codex/pull/23513), [#23582](https://github.com/openai/codex/pull/23582), [#23586](https://github.com/openai/codex/pull/23586), [#23596](https://github.com/openai/codex/pull/23596), [#23635](https://github.com/openai/codex/pull/23635), [#23636](https://github.com/openai/codex/pull/23636), [#23637](https://github.com/openai/codex/pull/23637), [#23638](https://github.com/openai/codex/pull/23638), [#23786](https://github.com/openai/codex/pull/23786))\n- Fixed Linux Python runtime wheel tags so glibc-based systems can install the runtime artifacts. ([#21812](https://github.com/openai/codex/pull/21812))\n- Improved release and CI reliability with package-builder tests, prebuilt resource packaging, DotSlash zstd handling, platform-sharded Rust tests, and Codex Linux release runners. ([#23760](https://github.com/openai/codex/pull/23760), [#23759](https://github.com/openai/codex/pull/23759), [#23752](https://github.com/openai/codex/pull/23752), [#23358](https://github.com/openai/codex/pull/23358), [#23761](https://github.com/openai/codex/pull/23761))\n\n## Changelog\n\nFull Changelog: [rust-v0.132.0...rust-v0.133.0](https://github.com/openai/codex/compare/rust-v0.132.0...rust-v0.133.0)\n\n- [#23343](https://github.com/openai/codex/pull/23343) codex: route global AGENTS reads through LOCAL\\_FS [@starr-openai](https://github.com/starr-openai)\n- [#22380](https://github.com/openai/codex/pull/22380) fix: default unknown tool schemas to empty schemas [@celia-oai](https://github.com/celia-oai)\n- [#23309](https://github.com/openai/codex/pull/23309) Add tool lifecycle extension contributor [@jif-oai](https://github.com/jif-oai)\n- [#23253](https://github.com/openai/codex/pull/23253) Reduce rust-ci-full Windows nextest timeout flakes [@starr-openai](https://github.com/starr-openai)\n- [#22878](https://github.com/openai/codex/pull/22878) Improve `codex remote-control` CLI UX [@owenlin0](https://github.com/owenlin0)\n- [#21812](https://github.com/openai/codex/pull/21812) Publish Linux runtime wheels with glibc-compatible tags [@aibrahim-oai](https://github.com/aibrahim-oai)\n- [#22709](https://github.com/openai/codex/pull/22709) \\[codex\\] Trim unused TurnContextItem fields [@pakrym-oai](https://github.com/pakrym-oai)\n- [#23353](https://github.com/openai/codex/pull/23353) Include plugin id in plugin MCP tool metadata [@mzeng-openai](https://github.com/mzeng-openai)\n- [#22728](https://github.com/openai/codex/pull/22728) \\[codex\\] Move pending input into input queue [@pakrym-oai](https://github.com/pakrym-oai)\n- [#23371](https://github.com/openai/codex/pull/23371) fix(tui): warn on unsupported iTerm2 pet versions [@fcoury-oai](https://github.com/fcoury-oai)\n- [#23376](https://github.com/openai/codex/pull/23376) \\[codex-analytics\\] preserve user thread source for exec threads [@marksteinbrick-oai](https://github.com/marksteinbrick-oai)\n- [#23360](https://github.com/openai/codex/pull/23360) app-server: use profile ids in v2 permission params [@bolinfest](https://github.com/bolinfest)\n- [#23384](https://github.com/openai/codex/pull/23384) \\[codex\\] Remove external websocket session resets [@pakrym-oai](https://github.com/pakrym-oai)\n- [#22721](https://github.com/openai/codex/pull/22721) cleanup: Remove skill env var dependency prompting [@xl-openai](https://github.com/xl-openai)\n- [#23389](https://github.com/openai/codex/pull/23389) Remove ToolSearch feature toggle [@sayan-oai](https://github.com/sayan-oai)\n- [#23080](https://github.com/openai/codex/pull/23080) \\[1 of 7\\] Add thread settings to UserInput [@etraut-openai](https://github.com/etraut-openai)\n- [#23081](https://github.com/openai/codex/pull/23081) \\[2 of 7\\] Remove UserInputWithTurnContext [@etraut-openai](https://github.com/etraut-openai)\n- [#23075](https://github.com/openai/codex/pull/23075) \\[3 of 7\\] Remove UserTurn [@etraut-openai](https://github.com/etraut-openai)\n- [#23396](https://github.com/openai/codex/pull/23396) \\[codex\\] Extract turn skill and plugin injections [@pakrym-oai](https://github.com/pakrym-oai)\n- [#23356](https://github.com/openai/codex/pull/23356) fix(plugins): keep version upgrades additive [@iceweasel-oai](https://github.com/iceweasel-oai)\n- [#22508](https://github.com/openai/codex/pull/22508) \\[5 of 7\\] Replace OverrideTurnContext with ThreadSettings [@etraut-openai](https://github.com/etraut-openai)\n- [#22086](https://github.com/openai/codex/pull/22086) CI: Customize v8 building [@cconger](https://github.com/cconger)\n- [#23390](https://github.com/openai/codex/pull/23390) Remove explicit connector tool undeferral [@sayan-oai](https://github.com/sayan-oai)\n- [#22928](https://github.com/openai/codex/pull/22928) core: expose permission profile picker metadata [@viyatb-oai](https://github.com/viyatb-oai)\n- [#23352](https://github.com/openai/codex/pull/23352) Preserve context baselines for full-history agent forks [@jif-oai](https://github.com/jif-oai)\n- [#23300](https://github.com/openai/codex/pull/23300) feat: dedicated goal DB [@jif-oai](https://github.com/jif-oai)\n- [#22835](https://github.com/openai/codex/pull/22835) Remove ToolsConfig from tool planning [@jif-oai](https://github.com/jif-oai)\n- [#22870](https://github.com/openai/codex/pull/22870) Add `body_after_prefix` auto-compact token limit scope [@jif-oai](https://github.com/jif-oai)\n- [#23144](https://github.com/openai/codex/pull/23144) Defer v1 multi-agent tools behind tool search [@jif-oai](https://github.com/jif-oai)\n- [#23409](https://github.com/openai/codex/pull/23409) \\[codex\\] Allow empty turn/start requests [@pakrym-oai](https://github.com/pakrym-oai)\n- [#23388](https://github.com/openai/codex/pull/23388) \\[codex\\] Move hook request plumbing into hook runtime [@pakrym-oai](https://github.com/pakrym-oai)\n- [#23405](https://github.com/openai/codex/pull/23405) \\[codex\\] Preserve steer input as user input [@pakrym-oai](https://github.com/pakrym-oai)\n- [#22914](https://github.com/openai/codex/pull/22914) \\[2 of 4\\] tui: route app and skill enablement through app server [@etraut-openai](https://github.com/etraut-openai)\n- [#23397](https://github.com/openai/codex/pull/23397) \\[codex\\] Make contextual user fragments dyn-renderable [@pakrym-oai](https://github.com/pakrym-oai)\n- [#23475](https://github.com/openai/codex/pull/23475) chore: namespace v1 sub-agent tools [@jif-oai](https://github.com/jif-oai)\n- [#23493](https://github.com/openai/codex/pull/23493) Make `deny` canonical for filesystem permission entries [@viyatb-oai](https://github.com/viyatb-oai)\n- [#22929](https://github.com/openai/codex/pull/22929) Harden CLI rate limit window labels [@ase-openai](https://github.com/ase-openai)\n- [#22782](https://github.com/openai/codex/pull/22782) Add SubagentStart hook [@abhinav-oai](https://github.com/abhinav-oai)\n- [#23513](https://github.com/openai/codex/pull/23513) build: add Codex package builder [@bolinfest](https://github.com/bolinfest)\n- [#23369](https://github.com/openai/codex/pull/23369) Make local environment optional in EnvironmentManager [@starr-openai](https://github.com/starr-openai)\n- [#23327](https://github.com/openai/codex/pull/23327) Refactor exec-server websocket pump [@starr-openai](https://github.com/starr-openai)\n- [#23536](https://github.com/openai/codex/pull/23536) fix(tui): preserve modified enter in plan questions [@fcoury-oai](https://github.com/fcoury-oai)\n- [#23400](https://github.com/openai/codex/pull/23400) Fix empty rollout path app-server handling [@wiltzius-openai](https://github.com/wiltzius-openai)\n- [#23551](https://github.com/openai/codex/pull/23551) Route local-only app-server gating through processors [@starr-openai](https://github.com/starr-openai)\n- [#23372](https://github.com/openai/codex/pull/23372) Split plugin install discovery into list and request tools [@mzeng-openai](https://github.com/mzeng-openai)\n- [#23516](https://github.com/openai/codex/pull/23516) fix: serialize unix app-server startup [@efrazer-oai](https://github.com/efrazer-oai)\n- [#22169](https://github.com/openai/codex/pull/22169) \\[codex\\] Honor role-defined spawn service tiers [@aibrahim-oai](https://github.com/aibrahim-oai)\n- [#23555](https://github.com/openai/codex/pull/23555) Add CUA requirements subsection for locked computer use [@adams-oai](https://github.com/adams-oai)\n- [#23538](https://github.com/openai/codex/pull/23538) Fix: TUI starting in wrong CWD [@canvrno-oai](https://github.com/canvrno-oai)\n- [#23526](https://github.com/openai/codex/pull/23526) build: fetch rg for Codex packages [@bolinfest](https://github.com/bolinfest)\n- [#23573](https://github.com/openai/codex/pull/23573) Remove unused ARC monitor path [@mzeng-openai](https://github.com/mzeng-openai)\n- [#23576](https://github.com/openai/codex/pull/23576) test: fix multi-agent service tier assertion [@bolinfest](https://github.com/bolinfest)\n- [#23541](https://github.com/openai/codex/pull/23541) build: default Codex package target and output [@bolinfest](https://github.com/bolinfest)\n- [#23358](https://github.com/openai/codex/pull/23358) Fan out rust-ci-full nextest by platform [@starr-openai](https://github.com/starr-openai)\n- [#23593](https://github.com/openai/codex/pull/23593) feat: expose codex-app-server version flag [@bolinfest](https://github.com/bolinfest)\n- [#23412](https://github.com/openai/codex/pull/23412) feat: add permission profile list api [@viyatb-oai](https://github.com/viyatb-oai)\n- [#23535](https://github.com/openai/codex/pull/23535) Move plugin and skill warmup into session startup [@aibrahim-oai](https://github.com/aibrahim-oai)\n- [#23231](https://github.com/openai/codex/pull/23231) Fix stale background terminal poll events [@etraut-openai](https://github.com/etraut-openai)\n- [#23564](https://github.com/openai/codex/pull/23564) \\[codex\\] Preserve raw code-mode exec output by default [@aibrahim-oai](https://github.com/aibrahim-oai)\n- [#23232](https://github.com/openai/codex/pull/23232) Warn on invalid UTF-8 in AGENTS.md files [@etraut-openai](https://github.com/etraut-openai)\n- [#23584](https://github.com/openai/codex/pull/23584) feat: Add vertical remote plugin collection support [@xl-openai](https://github.com/xl-openai)\n- [#23586](https://github.com/openai/codex/pull/23586) build: package prebuilt Codex entrypoints [@bolinfest](https://github.com/bolinfest)\n- [#23582](https://github.com/openai/codex/pull/23582) ci: build Codex package archives in release workflow [@bolinfest](https://github.com/bolinfest)\n- [#23596](https://github.com/openai/codex/pull/23596) runtime: detect Codex package layout [@bolinfest](https://github.com/bolinfest)\n- [#23500](https://github.com/openai/codex/pull/23500) add encryptedcontent to functioncalloutput [@sayan-oai](https://github.com/sayan-oai)\n- [#23633](https://github.com/openai/codex/pull/23633) Migrate exec-server remote registration to environments [@richardopenai](https://github.com/richardopenai)\n- [#23451](https://github.com/openai/codex/pull/23451) Add timeout for remote compaction requests [@jif-oai](https://github.com/jif-oai)\n- [#23667](https://github.com/openai/codex/pull/23667) feat: rename 1 [@jif-oai](https://github.com/jif-oai)\n- [#23669](https://github.com/openai/codex/pull/23669) feat: rename 3 [@jif-oai](https://github.com/jif-oai)\n- [#23668](https://github.com/openai/codex/pull/23668) feat: rename 2 [@jif-oai](https://github.com/jif-oai)\n- [#23675](https://github.com/openai/codex/pull/23675) fix: main [@jif-oai](https://github.com/jif-oai)\n- [#23685](https://github.com/openai/codex/pull/23685) feat: wire goal extension tools to the dedicated goal store [@jif-oai](https://github.com/jif-oai)\n- [#23690](https://github.com/openai/codex/pull/23690) feat: async approval contrib [@jif-oai](https://github.com/jif-oai)\n- [#23692](https://github.com/openai/codex/pull/23692) feat: async turn item process [@jif-oai](https://github.com/jif-oai)\n- [#23688](https://github.com/openai/codex/pull/23688) feat: expose turn-start metadata to extensions [@jif-oai](https://github.com/jif-oai)\n- [#23605](https://github.com/openai/codex/pull/23605) \\[codex\\] Hide deferred tools from code mode prompt [@pakrym-oai](https://github.com/pakrym-oai)\n- [#23634](https://github.com/openai/codex/pull/23634) runtime: use install context for bundled bwrap [@bolinfest](https://github.com/bolinfest)\n- [#23635](https://github.com/openai/codex/pull/23635) release: publish Codex package archive checksums [@bolinfest](https://github.com/bolinfest)\n- [#23592](https://github.com/openai/codex/pull/23592) feat: Add btw alias for side slash command [@anp-oai](https://github.com/anp-oai)\n- [#23696](https://github.com/openai/codex/pull/23696) feat: account active goal progress in the goal extension [@jif-oai](https://github.com/jif-oai)\n- [#23176](https://github.com/openai/codex/pull/23176) \\[2 of 2\\] Start fresh TUI thread in background [@etraut-openai](https://github.com/etraut-openai)\n- [#23578](https://github.com/openai/codex/pull/23578) fix(app-server): speed up shutdown [@fcoury-oai](https://github.com/fcoury-oai)\n- [#22896](https://github.com/openai/codex/pull/22896) windows-sandbox: add resolved permissions helper [@bolinfest](https://github.com/bolinfest)\n- [#23502](https://github.com/openai/codex/pull/23502) Add thread/settings/update app-server API [@etraut-openai](https://github.com/etraut-openai)\n- [#23507](https://github.com/openai/codex/pull/23507) Sync TUI thread settings through app server [@etraut-openai](https://github.com/etraut-openai)\n- [#23666](https://github.com/openai/codex/pull/23666) feat: add turn\\_id and truncation\\_policy to extension tool calls [@jif-oai](https://github.com/jif-oai)\n- [#23636](https://github.com/openai/codex/pull/23636) install: consume Codex package archives [@bolinfest](https://github.com/bolinfest)\n- [#23717](https://github.com/openai/codex/pull/23717) \\[codex\\] Preserve failed goal accounting flushes [@jif-oai](https://github.com/jif-oai)\n- [#23655](https://github.com/openai/codex/pull/23655) add standalone websearch api client [@sayan-oai](https://github.com/sayan-oai)\n- [#23724](https://github.com/openai/codex/pull/23724) Fix thread settings clippy failure [@etraut-openai](https://github.com/etraut-openai)\n- [#23637](https://github.com/openai/codex/pull/23637) npm: ship platform packages in Codex package layout [@bolinfest](https://github.com/bolinfest)\n- [#23729](https://github.com/openai/codex/pull/23729) fix(config): resolve cloud requirements deny-read globs [@viyatb-oai](https://github.com/viyatb-oai)\n- [#23638](https://github.com/openai/codex/pull/23638) dotslash: publish Codex entrypoints from package archives [@bolinfest](https://github.com/bolinfest)\n- [#22918](https://github.com/openai/codex/pull/22918) windows-sandbox: send permission profiles to elevated runner [@bolinfest](https://github.com/bolinfest)\n- [#23735](https://github.com/openai/codex/pull/23735) windows-sandbox: share bundled helper lookup [@bolinfest](https://github.com/bolinfest)\n- [#18868](https://github.com/openai/codex/pull/18868) Add MITM hook config model [@evawong-oai](https://github.com/evawong-oai)\n- [#22270](https://github.com/openai/codex/pull/22270) feat(permissions): resolve permission profile inheritance [@viyatb-oai](https://github.com/viyatb-oai)\n- [#23719](https://github.com/openai/codex/pull/23719) cli: add strict config to exec-server [@bolinfest](https://github.com/bolinfest)\n- [#23542](https://github.com/openai/codex/pull/23542) \\[skills\\] Create a personal update flow for plugin creator [@caseychow-oai](https://github.com/caseychow-oai)\n- [#21272](https://github.com/openai/codex/pull/21272) Support compact SessionStart hooks [@abhinav-oai](https://github.com/abhinav-oai)\n- [#20659](https://github.com/openai/codex/pull/20659) Wire MITM hooks into runtime enforcement [@evawong-oai](https://github.com/evawong-oai)\n- [#23752](https://github.com/openai/codex/pull/23752) release: use DotSlash zstd for package archives [@bolinfest](https://github.com/bolinfest)\n- [#22923](https://github.com/openai/codex/pull/22923) windows-sandbox: drive write roots from resolved permissions [@bolinfest](https://github.com/bolinfest)\n- [#23761](https://github.com/openai/codex/pull/23761) chore: use Codex Linux runners for Rust releases [@bolinfest](https://github.com/bolinfest)\n- [#23759](https://github.com/openai/codex/pull/23759) release: package prebuilt resource binaries [@bolinfest](https://github.com/bolinfest)\n- [#23167](https://github.com/openai/codex/pull/23167) windows-sandbox: feed setup from resolved permissions [@bolinfest](https://github.com/bolinfest)\n- [#22931](https://github.com/openai/codex/pull/22931) core: refresh active permission profiles at runtime [@viyatb-oai](https://github.com/viyatb-oai)\n- [#22873](https://github.com/openai/codex/pull/22873) Add SubagentStop hook [@abhinav-oai](https://github.com/abhinav-oai)\n- [#23727](https://github.com/openai/codex/pull/23727) feat(plugins): tabulate plugin list output [@caseychow-oai](https://github.com/caseychow-oai)\n- [#23732](https://github.com/openai/codex/pull/23732) Make goals feature on by default and no longer experimental [@etraut-openai](https://github.com/etraut-openai)\n- [#23537](https://github.com/openai/codex/pull/23537) Honor client-resolved service tier defaults [@shijie-oai](https://github.com/shijie-oai)\n- [#23771](https://github.com/openai/codex/pull/23771) \\[codex\\] Fix realtime v1 websocket compatibility [@guinness-oai](https://github.com/guinness-oai)\n- [#23764](https://github.com/openai/codex/pull/23764) Remove Windows sandbox resource stamping [@iceweasel-oai](https://github.com/iceweasel-oai)\n- [#23730](https://github.com/openai/codex/pull/23730) \\[codex\\] List marketplaces considered by plugin discovery [@caseychow-oai](https://github.com/caseychow-oai)\n- [#23760](https://github.com/openai/codex/pull/23760) ci: run Codex package builder tests [@bolinfest](https://github.com/bolinfest)\n- [#23737](https://github.com/openai/codex/pull/23737) \\[codex\\] Add plugin id to MCP tool call items [@mzeng-openai](https://github.com/mzeng-openai)\n- [#18240](https://github.com/openai/codex/pull/18240) Use named MITM permissions config [@evawong-oai](https://github.com/evawong-oai)\n- [#23774](https://github.com/openai/codex/pull/23774) \\[codex\\] Reject read-only fallback with approvals disabled [@viyatb-oai](https://github.com/viyatb-oai)\n- [#23714](https://github.com/openai/codex/pull/23714) windows-sandbox: add profile-native elevated APIs [@bolinfest](https://github.com/bolinfest)\n- [#23433](https://github.com/openai/codex/pull/23433) feat: support managed permission profiles in requirements.toml [@viyatb-oai](https://github.com/viyatb-oai)\n- [#23715](https://github.com/openai/codex/pull/23715) core: pass permission profiles to Windows runner [@bolinfest](https://github.com/bolinfest)\n- [#23786](https://github.com/openai/codex/pull/23786) sdk: launch packaged Codex runtimes [@bolinfest](https://github.com/bolinfest)","publishedAt":"2026-05-21T00:00:00.000Z","url":"https://developers.openai.com/codex/changelog/#github-release-327109656","media":[],"prerelease":false,"source":{"slug":"openai-codex-changelog","name":"Codex Changelog","type":"feed"},"org":{"slug":"openai","name":"OpenAI"},"product":{"slug":"codex","name":"Codex"},"coverageCount":0},{"id":"rel_sc85xP00Ev5P9ievD6z8N","version":null,"type":"feature","title":"Improvements to Cursor Automations","summary":"This release brings [Cursor Automations](https://cursor.com/docs/cloud-agent/automations) to the Agents Window and introduces the ability to configure...","titleGenerated":null,"titleShort":null,"content":"This release brings [Cursor Automations](https://cursor.com/docs/cloud-agent/automations) to the Agents Window and introduces the ability to configure automations with multiple attached repos or no repos at all.\n\nFor the next 7 days, all agent runs for newly created automations are 50% off.\n\n## Automations in the Agents Window\n\nCursor Automations are now available in the [Agents Window](https://cursor.com/blog/cursor-3), in addition to cursor.com/automations. Create and manage your automations in the same workspace as your agents.\n\n## Multi-repo automations\n\nA lot of engineering work spans more than one codebase. You can now attach multiple repos to an automation so agents reason across all required context and work across repos to deliver, test, and verify tasks.\n\n## No-repo automations\n\nMany useful automations exist apart from code, where agents monitor your tools and act on key signals. You can now create automations without an attached repository.\n\nWe've added five new templates for no-repo automations to the [Cursor Marketplace](https://cursor.com/marketplace) to help you get started:\n\n- [\\*\\*Slack digest agent\\*\\*](https://cursor.com/marketplace/automations/slack-digest-agent): Summarizes unread DMs and key Slack channels every morning and prioritizes them by importance\n- [\\*\\*Product analytics agent\\*\\*](https://cursor.com/marketplace/automations/product-analytics-agent): Delivers a weekly digest of key metrics from your data warehouse like Databricks\n- [\\*\\*Product FAQ agent\\*\\*](https://cursor.com/marketplace/automations/product-faq-agent): Watches a Slack channel for questions and writes a first response based on docs, codebase context, and past threads\n- [\\*\\*Product finance agent\\*\\*](https://cursor.com/marketplace/automations/product-finance-agent): Pulls financial data from a billing provider like Stripe for recurring revenue reports\n- [\\*\\*Customer health agent\\*\\*](https://cursor.com/marketplace/automations/customer-health-monitoring-agent): Monitors key systems like Granola, Slack, and Databricks and flags accounts where health signals are shifting\n\nRead our [docs](https://cursor.com/docs/cloud-agent/automations) to learn more.","publishedAt":"2026-05-20T00:00:00.000Z","url":"https://cursor.com/changelog/05-20-26","media":[],"prerelease":false,"source":{"slug":"cursor-changelog","name":"Cursor Changelog","type":"feed"},"org":{"slug":"cursor","name":"Cursor"},"product":null,"coverageCount":0},{"id":"rel_TgNY23Sb9ORLF9ab-SYT9","version":"0.132.0","type":"feature","title":"Codex CLI Release: 0.132.0","summary":"The Python SDK now supports first-class authentication with API key login, ChatGPT browser and device-code flows, account inspection, and logout APIs. TUI startup is faster because terminal capability probes are now batched. Goal continuations now stop when hitting usage limits instead of looping and burning tokens, and remote sessions keep websocket connections alive.","titleGenerated":"Codex CLI 0.132.0 adds first-class Python SDK authentication and faster TUI startup","titleShort":"Python SDK gains first-class auth; TUI startup faster","content":"## New Features\n\n- The Python SDK now supports first-class authentication, including API key login, ChatGPT browser and device-code flows, account inspection, and logout APIs. ([#23093](https://github.com/openai/codex/pull/23093))\n- Python turn APIs are easier to use for text-only workflows: you can pass a plain string as input, and handle-based runs now return a richer `TurnResult` with collected items, timing, and usage data. ([#23151](https://github.com/openai/codex/pull/23151), [#23162](https://github.com/openai/codex/pull/23162))\n- `codex exec resume` now accepts `--output-schema`, so resumed automations can keep session context while still enforcing structured JSON output. ([#23123](https://github.com/openai/codex/pull/23123))\n- TUI startup is faster because terminal capability probes are now batched instead of waiting on several serial checks before the first interactive frame. ([#23175](https://github.com/openai/codex/pull/23175))\n- Remote executor registration can now use standard Codex auth instead of a separate registry credential flow. ([#22769](https://github.com/openai/codex/pull/22769))\n- App-server turns can preserve requested image fidelity, including original-resolution local images, across user inputs and image-producing tools. ([#20693](https://github.com/openai/codex/pull/20693))\n\n## Bug Fixes\n\n- Goal continuations now stop when they hit usage limits or a repeated blocker instead of looping and burning more tokens, and completion responses phrase usage more naturally. ([#23094](https://github.com/openai/codex/pull/23094), [#22907](https://github.com/openai/codex/pull/22907))\n- The session picker is easier to trust: renamed threads now show `name (thread-id)` in resume hints, and pasted text works in the picker search box. ([#23234](https://github.com/openai/codex/pull/23234), [#23338](https://github.com/openai/codex/pull/23338))\n- Multi-session TUI flows are more reliable: in-progress MCP calls stay marked as active during replay, and elicitation replies are sent back to the thread that requested them. ([#23236](https://github.com/openai/codex/pull/23236), [#23241](https://github.com/openai/codex/pull/23241))\n- Remote sessions now keep websocket connections alive and show repo-relative diff paths again instead of `/tmp/...`\\-prefixed paths. ([#23226](https://github.com/openai/codex/pull/23226), [#23261](https://github.com/openai/codex/pull/23261))\n- Windows installs are more robust: `codex doctor` now detects npm-managed installs correctly, and MSVC release binaries no longer depend on separately installed VC++ runtime DLLs. ([#22967](https://github.com/openai/codex/pull/22967), [#22905](https://github.com/openai/codex/pull/22905))\n- TUI polish fixes include immediate shutdown feedback on exit, hiding the ChatGPT usage link for non-OpenAI providers, and keeping a cleared Fast tier from reappearing after side-thread resume. ([#23323](https://github.com/openai/codex/pull/23323), [#23127](https://github.com/openai/codex/pull/23127), [#23121](https://github.com/openai/codex/pull/23121))\n\n## Documentation\n\n- The Python SDK docs, FAQ, and examples were refreshed around the new auth flow and turn APIs, with clearer setup guidance and simpler text-only examples. ([#22941](https://github.com/openai/codex/pull/22941), [#23093](https://github.com/openai/codex/pull/23093), [#23151](https://github.com/openai/codex/pull/23151), [#23162](https://github.com/openai/codex/pull/23162))\n\n## Chores\n\n- Memory summaries are now versioned and rebuilt when the stored format is stale, which should keep long-lived memory context leaner and more predictable. ([#23148](https://github.com/openai/codex/pull/23148))\n\n## Changelog\n\nFull Changelog: [rust-v0.131.0...rust-v0.132.0](https://github.com/openai/codex/compare/rust-v0.131.0...rust-v0.132.0)\n\n- [#20693](https://github.com/openai/codex/pull/20693) Preserve image detail in app-server inputs [@fjord-oai](https://github.com/fjord-oai)\n- [#22891](https://github.com/openai/codex/pull/22891) tui: pass active permission profiles through app commands [@bolinfest](https://github.com/bolinfest)\n- [#22924](https://github.com/openai/codex/pull/22924) app-server-protocol: remove PermissionProfile from API [@bolinfest](https://github.com/bolinfest)\n- [#22941](https://github.com/openai/codex/pull/22941) \\[codex\\] Refine Python SDK user-facing docs [@aibrahim-oai](https://github.com/aibrahim-oai)\n- [#22967](https://github.com/openai/codex/pull/22967) Fix Windows doctor npm root probe [@etraut-openai](https://github.com/etraut-openai)\n- [#22920](https://github.com/openai/codex/pull/22920) core: set permission profiles from snapshots [@bolinfest](https://github.com/bolinfest)\n- [#22939](https://github.com/openai/codex/pull/22939) \\[codex\\] Split Python SDK helper logic [@aibrahim-oai](https://github.com/aibrahim-oai)\n- [#22907](https://github.com/openai/codex/pull/22907) Improve goal completion usage reporting [@etraut-openai](https://github.com/etraut-openai)\n- [#23030](https://github.com/openai/codex/pull/23030) test: construct permission profiles directly [@bolinfest](https://github.com/bolinfest)\n- [#22769](https://github.com/openai/codex/pull/22769) exec-server: support auth-backed remote executor registration [@miz-openai](https://github.com/miz-openai)\n- [#22946](https://github.com/openai/codex/pull/22946) \\[codex\\] preserve MCP result meta in McpToolCallItemResult [@miaolin-oai](https://github.com/miaolin-oai)\n- [#23069](https://github.com/openai/codex/pull/23069) multiagent: trim model-visible description, cap to 5 models [@sayan-oai](https://github.com/sayan-oai)\n- [#22913](https://github.com/openai/codex/pull/22913) \\[1 of 4\\] tui: route primary settings writes through app server [@etraut-openai](https://github.com/etraut-openai)\n- [#23093](https://github.com/openai/codex/pull/23093) sdk/python: add first-class login support [@aibrahim-oai](https://github.com/aibrahim-oai)\n- [#23151](https://github.com/openai/codex/pull/23151) \\[codex\\] Return TurnResult from Python turn handles [@aibrahim-oai](https://github.com/aibrahim-oai)\n- [#23147](https://github.com/openai/codex/pull/23147) Make multi-agent v2 tool namespace configurable [@jif-oai](https://github.com/jif-oai)\n- [#23036](https://github.com/openai/codex/pull/23036) test: reduce core sandbox policy test setup [@bolinfest](https://github.com/bolinfest)\n- [#23162](https://github.com/openai/codex/pull/23162) \\[codex\\] Accept string input for Python turns [@aibrahim-oai](https://github.com/aibrahim-oai)\n- [#23226](https://github.com/openai/codex/pull/23226) Add exec-server websocket keepalive [@starr-openai](https://github.com/starr-openai)\n- [#23148](https://github.com/openai/codex/pull/23148) Densify and version memory summaries [@jif-oai](https://github.com/jif-oai)\n- [#22448](https://github.com/openai/codex/pull/22448) \\[codex\\] Add installed-plugin mention API [@xli-oai](https://github.com/xli-oai)\n- [#23288](https://github.com/openai/codex/pull/23288) chore: goal ext skeleton [@jif-oai](https://github.com/jif-oai)\n- [#23291](https://github.com/openai/codex/pull/23291) Make extension lifecycle hooks async [@jif-oai](https://github.com/jif-oai)\n- [#23293](https://github.com/openai/codex/pull/23293) feat: add extension event sink capability [@jif-oai](https://github.com/jif-oai)\n- [#23295](https://github.com/openai/codex/pull/23295) chore: isolate thread goal storage behind GoalStore [@jif-oai](https://github.com/jif-oai)\n- [#23301](https://github.com/openai/codex/pull/23301) chore: goal resumed metrics [@jif-oai](https://github.com/jif-oai)\n- [#23305](https://github.com/openai/codex/pull/23305) chore: make token usage async [@jif-oai](https://github.com/jif-oai)\n- [#23306](https://github.com/openai/codex/pull/23306) Emit goal update events from goal extension tools [@jif-oai](https://github.com/jif-oai)\n- [#23121](https://github.com/openai/codex/pull/23121) tui: keep cleared Fast tier from reappearing after side-thread resume [@etraut-openai](https://github.com/etraut-openai)\n- [#23123](https://github.com/openai/codex/pull/23123) Support --output-schema for exec resume [@etraut-openai](https://github.com/etraut-openai)\n- [#23128](https://github.com/openai/codex/pull/23128) Fix TUI stream cleanup after turn errors [@etraut-openai](https://github.com/etraut-openai)\n- [#23127](https://github.com/openai/codex/pull/23127) Hide ChatGPT usage link for non-OpenAI status [@etraut-openai](https://github.com/etraut-openai)\n- [#23175](https://github.com/openai/codex/pull/23175) \\[1 of 2\\] Optimize TUI startup terminal probes [@etraut-openai](https://github.com/etraut-openai)\n- [#22706](https://github.com/openai/codex/pull/22706) \\[codex\\] Remove legacy shell output formatting paths [@pakrym-oai](https://github.com/pakrym-oai)\n- [#23332](https://github.com/openai/codex/pull/23332) nit: read prompt [@jif-oai](https://github.com/jif-oai)\n- [#22905](https://github.com/openai/codex/pull/22905) windows: link MSVC release binaries with static CRT [@iceweasel-oai](https://github.com/iceweasel-oai)\n- [#23323](https://github.com/openai/codex/pull/23323) fix(tui): show shutdown feedback on exit [@fcoury-oai](https://github.com/fcoury-oai)\n- [#23261](https://github.com/openai/codex/pull/23261) Fix remote turn diff display roots [@starr-openai](https://github.com/starr-openai)\n- [#22569](https://github.com/openai/codex/pull/22569) Simplify legacy Windows sandbox ACL persistence [@iceweasel-oai](https://github.com/iceweasel-oai)\n- [#23273](https://github.com/openai/codex/pull/23273) Upload rust full CI JUnit reports [@starr-openai](https://github.com/starr-openai)\n- [#22893](https://github.com/openai/codex/pull/22893) fix: harden plugin creator sharing validation [@efrazer-oai](https://github.com/efrazer-oai)\n- [#23094](https://github.com/openai/codex/pull/23094) goal: pause continuation loops on usage limits and blockers [@etraut-openai](https://github.com/etraut-openai)\n- [#23234](https://github.com/openai/codex/pull/23234) Clarify resume hints for renamed threads [@etraut-openai](https://github.com/etraut-openai)\n- [#23241](https://github.com/openai/codex/pull/23241) TUI: route elicitation responses to request thread [@etraut-openai](https://github.com/etraut-openai)\n- [#23236](https://github.com/openai/codex/pull/23236) TUI: replay in-progress MCP calls as started [@etraut-openai](https://github.com/etraut-openai)\n- [#23088](https://github.com/openai/codex/pull/23088) goals: keep pause transitions explicit [@etraut-openai](https://github.com/etraut-openai)\n- [#23338](https://github.com/openai/codex/pull/23338) feat(tui): handle paste in session picker [@fcoury-oai](https://github.com/fcoury-oai)\n- [#23335](https://github.com/openai/codex/pull/23335) feat(app-server): add optional thread\\_id to experimentalFeature/list [@owenlin0](https://github.com/owenlin0)","publishedAt":"2026-05-20T00:00:00.000Z","url":"https://developers.openai.com/codex/changelog/#github-release-325545332","media":[],"prerelease":false,"source":{"slug":"openai-codex-changelog","name":"Codex Changelog","type":"feed"},"org":{"slug":"openai","name":"OpenAI"},"product":{"slug":"codex","name":"Codex"},"coverageCount":0},{"id":"rel_wDsKCiz48sDRzIbaM_rnw","version":"v2.1.145","type":"feature","title":"v2.1.145","summary":"Fixed a permission-prompt bypass where bare variable assignments to non-allowlisted environment variables in Bash commands were auto-approved. Also fixed Agent Teams teammates with non-ASCII names failing API calls due to invalid header encoding. Plus session list JSON export, plugin discovery improvements, and dozens of smaller fixes across MCP commands, UI rendering, and cross-platform support.","titleGenerated":"Claude Code v2.1.145 fixes permission-prompt bypass and permission header encoding","titleShort":"Permission-prompt bypass closed; non-ASCII names now work","content":"## What's changed\n\n- Added `claude agents --json` to list live Claude sessions as JSON for scripting (tmux-resurrect, status bars, session pickers)\n- Added `agent_id` and `parent_agent_id` attributes to `claude_code.tool` OTEL spans, and fixed trace parenting so background subagent spans nest under the dispatching Agent tool span\n- Status line JSON input now includes GitHub repo and PR information when detected\n- `/plugin` Discover and Browse screens now show a plugin's commands, agents, skills, hooks, and MCP/LSP servers before installation\n- `claude agents` terminal tab title now shows the awaiting-input count so an alt-tabbed window tells you when an agent needs attention\n- Slash command and @-mention suggestion list now supports mouse hover and click in fullscreen mode\n- Stop and SubagentStop hook input now includes `background_tasks` and `session_crons` fields\n- Fixed a permission-prompt bypass where bare variable assignments to non-allowlisted environment variables in Bash commands were auto-approved\n- Fixed MCP prompt slash commands showing raw server validation errors when a required argument is omitted — the error now names the missing argument and shows expected usage\n- Fixed the spinner and elapsed-time display freezing until a keypress after the terminal was resized or refocused\n- Fixed the cross-project resume hint failing in default Windows PowerShell 5.1 — Windows now uses `;` as the command separator\n- Fixed voice push-to-talk not working in the agent view's reply pane\n- Fixed task lists rendering in random order when several tasks are created at once\n- Fixed stale \"Failed to install Anthropic marketplace\" banner showing when the marketplace is already installed\n- Fixed the PR badge in the footer not updating immediately after `gh pr create` and other PR-state-changing commands run in-session\n- Fixed Agent Teams teammates with non-ASCII names failing every API call due to invalid header encoding\n- Fixed `/review` using a deprecated `projectCards` GraphQL query that errored on repos with Classic Projects\n- Fixed `claude plugin validate` not flagging `skills:` entries that point at a file instead of a directory — the error now suggests the parent directory\n- Fixed an infinite loop where a skill using `context: fork` could repeatedly re-invoke itself instead of running\n- Improved the Read tool to return a truncated first page with a \"PARTIAL view\" notice instead of a hard error when a whole-file read exceeds the token limit\n","publishedAt":"2026-05-19T21:31:07.000Z","url":"https://code.claude.com/docs/en/changelog#2-1-145","media":[],"prerelease":false,"source":{"slug":"claude-code","name":"Claude Code","type":"scrape"},"org":{"slug":"anthropic","name":"Anthropic"},"product":{"slug":"claude-code","name":"Claude Code"},"coverageCount":0},{"id":"rel_yVTX64BUt1JAMOcO0rT09","version":"v2.1.144","type":"feature","title":"v2.1.144","summary":"Fixed macOS background sessions crashing with \"exit 1 before init\" when the project lives under a Full Disk Access-protected folder, and progressive terminal display corruption (stale or garbled glyphs) that could appear in very long sessions. Also fixed startup hanging up to 75s when api.anthropic.com is unreachable, garbled terminal output after missed window-resize events, and dozens of MCP, tool, and session management issues.","titleGenerated":"Claude Code v2.1.144 fixes background session crashes and terminal rendering corruption","titleShort":"Background sessions no longer crash on Full Disk Access folders; terminal rendering stabilized","content":"## What's changed\n\n- Added `/resume` support for background sessions — sessions started via `claude --bg` or agent view now appear alongside interactive ones, marked with `bg`\n- Added elapsed duration to background subagent completion notifications (e.g. \"Agent completed · 3h 2m 5s\")\n- The `/plugin` browse and discover panes now show when a plugin was last updated\n- `/model` now changes the model for the current session only; press `d` in the model picker to set a default for new sessions\n- Renamed \"extra usage\" to \"usage credits\" across CLI copy; `/extra-usage` is now `/usage-credits` (old name still works)\n- Fixed startup hanging up to 75s when `api.anthropic.com` is unreachable (captive portal, firewall, VPN issues) — side-channel API calls now time out after 15s\n- Fixed garbled terminal output after a missed window-resize event (e.g. dragging a VS Code split-pane divider) — now self-heals on the next frame instead of requiring Ctrl+L\n- Fixed progressive terminal display corruption (stale/garbled glyphs) that could appear in very long sessions and only cleared on terminal resize or restart\n- Reduced terminal rendering glitches in VS Code by reducing spinner animation color count\n- Fixed macOS background sessions crashing with \"exit 1 before init\" when the project lives under a Full Disk Access-protected folder (regression in 2.1.143)\n- Fixed an unrecoverable conversation when reading a file whose image extension doesn't match its contents (e.g. HTML saved as .png) — now falls back to text\n- Fewer spurious tool errors during search: `head`/`tail` file views now satisfy the read-before-edit check, and a \"no matches\" result (exit code 1) from `egrep`, `fgrep`, `git grep`, or `git diff` is no longer reported as a command failure\n- Fixed `/branch` failing with \"No conversation to branch\" after entering a worktree or in some background sessions\n- Fixed pressing Escape in the AskUserQuestion notes field aborting the turn instead of returning to answer selection\n- Fixed model selection not applying when changed via the IDE model picker or `applyFlagSettings` after startup\n- Resumed sessions now keep the model they were using instead of picking up another session's `/model` choice\n- Fixed Bedrock and Vertex users unable to select \"Opus (1M context)\" from the `/model` picker (regression in v2.1.129)\n- Fixed remote-session login failing with \"Can't access this organization\" for users with `forceLoginMethod` and `forceLoginOrgUUID` set\n- Fixed MCP servers with paginated `tools/list` responses only returning the first page, silently dropping tools\n- Fixed MCP images with unsupported MIME types (e.g. SVG) breaking the conversation — now saved to disk and referenced in the tool result\n- Fixed file descriptor exhaustion when a build runs inside a skill directory — non-`.md` files no longer trigger skill reloads\n- Fixed session title being generated from plugin monitor output instead of the user's first prompt\n- Fixed Skill tool failing with permission error in headless mode (regression in v2.1.141)\n- Fixed plugins enabled in your own settings showing \"not cached\" errors after first load on a fresh machine; plugins enabled only by a project's `.claude/settings.json` now show an actionable `claude plugin install` hint\n- Fixed `claude mcp list` silently reporting no servers when `.mcp.json` can't be parsed (e.g. using VS Code's `\"servers\"` key instead of `\"mcpServers\"`) — now shows configuration errors\n- Fixed background side-queries on custom `ANTHROPIC_BASE_URL` setups and Bedrock Mantle not using Haiku — now falls back correctly when a first-party API key is configured or no Haiku model is set\n- Fixed scrolling in attached background sessions on Windows — PgUp/PgDn, mouse wheel, and Ctrl+O transcript navigation now work\n- Fixed a crash when closing the terminal while attached to a background session\n- Fixed `! ` exec sessions not responding to Ctrl+C while attached — now interrupts the running command\n- Fixed agent view shell-command rows lingering under Working after completion, and pressing Enter on a completed row re-running the command after its output expired\n- Fixed on Windows, pressing ← in `claude agents` leaving the list unresponsive to keyboard input\n- Fixed ghost characters at the left edge when switching panes in Agent View on Windows Terminal with CJK content\n- `/bg` and `←`-detach now preserve directories added via `/add-dir`\n- Fixed Edit/Write refusing with \"background session hasn't isolated its changes yet\" right after detaching a session that was already editing in place\n- Fixed `claude respawn ` on a stopped background session showing \"stopped\" instead of running\n- Fixed `/resume` picker not showing sessions forked from a background session\n- Fixed opening a session from `claude agents` or running `claude logs ` hanging when the background service is unresponsive — now times out after 10s with a recovery hint\n- Fixed background Bash tasks spawned by subagents staying \"Running\" in SDK task panels after the process exits\n- Fixed completed or stopped background sessions briefly failing to wake being permanently marked as a startup crash\n- Fixed markdown links in `claude agents` attached sessions rendering as plain text instead of clickable hyperlinks\n- Fixed custom `spinnerVerbs` applying to the post-turn duration message — past-tense built-ins like \"Worked for 5s\" are restored there\n- `claude agents` / `--bg` rejection messages now name the specific gate (non-TTY, env var, or setting) instead of a generic message\n- `claude --bg --name