Who contributed to OpenAI on July 22, 2026?

3 developers shipped this update, including rhan-oai, mandal-OAI, and copyberry.

What were the notable OpenAI updates?

Fix Windows TUI navigation key handling, Terminate Windows process trees with job objects, and Configure Codex Auto Review model metadata (#34687).

OpenAI

Codex, the SDKs, and the engine behind ChatGPT

Pick a date

Topics: AI / ML Full archive →

The Wire · Showcase

CODEX FIXES WINDOWS EXECUTION BUGS WHILE TIGHTENING SECURITY REVIEW GATES

By RepoJournal · Filed 06:59 UTC on July 22, 2026 · About OpenAI

3 people shipped this

copyberry @copyberry 3 cited

rhan-oai @rhan-oai 1 cited

mandal-OAI @mandal-OAI 1 cited

Codex shipped critical fixes for Windows process management and terminal navigation that were silently breaking execution sessions on Windows machines.

In brief

Codex now terminates Windows process trees using job objects instead of killing parent PIDs. Stops orphaned child processes from hanging around. should've been the default.
Fixed Windows TUI navigation key handling in Codex. Arrow keys and vim bindings were eating input in the wrong places. small bug, annoying fix.
Codex Auto Review model metadata is now configurable. gates review quality by rollout lineage. keeps experimental configs from leaking into prod.
Codex sanitizes skill names in injection metrics. stops parsing errors when names have special chars. boring security fix that matters.

Two production-grade Windows bugs landed overnight that explain intermittent failures developers have been seeing in TUI navigation and background process handling. The navigation fix [1] addresses a deeper issue where Crossterm's Windows backend was receiving literal escape bytes instead of proper input records when virtual terminal input mode got restored, making arrow keys and other controls unreliable. The process termination fix [2] ensures child processes actually die when a session is cancelled or times out on Windows, while still preserving background descendants on normal exit. These aren't edge cases. In parallel, Codex tightened its auto-review pipeline with new security policy defaults [3] that configure permission gates for `danger_full_access`, `workspace_write`, and `read_only` contexts. The team also fixed a metric pollution issue [4] where skill names with special characters were breaking observability, and shipped support for paginated model context across rollout lineages [5] to handle more complex session state loading.

Action items

→ If you're running Codex on Windows, pull the latest immediately. Navigation and process termination are now reliable. openai/codex [immediate]
→ Review your Codex auto-review configuration to ensure security policies match your permission model. openai/codex [plan]
→ Monitor metrics dashboards for skill injection counters. Tag values are now properly sanitized. openai/codex [monitor]

References

[1] Fix Windows TUI navigation key handling ↗ openai/codex
[2] Terminate Windows process trees with job objects ↗ openai/codex
[3] Configure Codex Auto Review model metadata (#34687) openai/codex
[4] Sanitize skill names in injection metrics (#34601) openai/codex
[5] Load paginated model context across rollout lineages ↗ openai/codex

Quick answers

What shipped in OpenAI on July 22, 2026?: Codex shipped critical fixes for Windows process management and terminal navigation that were silently breaking execution sessions on Windows machines. In total, 50 commits, 50 pull requests, and 4 releases landed.
Who contributed to OpenAI on July 22, 2026?: 3 developers shipped this update, including rhan-oai, mandal-OAI, and copyberry.
What were the notable OpenAI updates?: Fix Windows TUI navigation key handling, Terminate Windows process trees with job objects, and Configure Codex Auto Review model metadata (#34687).

CODEX TIGHTENS MEMORY HANDLING AND AUDIO PRESERVATION ACROSS THREADS

Codex shipped five structural fixes overnight that close gaps in memory reconciliation, audio attachment handling, and Windows permission checks that could silently corrupt state in production.

ai 105 shipped 1-min read

OpenAI 2 days ago

CODEX FIXES RACE CONDITIONS AND STABILIZES STREAMING PIPELINE

Codex shipped five architectural fixes overnight that eliminate liveness races in side conversations, stabilize the Markdown streaming stack, and cut TUI layout time significantly.

ai 34 shipped 1-min read

OpenAI 3 days ago

CODEX HOTFIX CORRECTS GPT-5.6 CONTEXT WINDOWS AFTER BUNDLED INSTRUCTION REFRESH

Codex 0.144.6 shipped overnight to fix context window mismatches across GPT-5.6 Sol, Terra, and Luna after bundled instruction metadata got out of sync with the models themselves.

ai 18 shipped 1-min read

OpenAI 4 days ago

CODEX LOCKS DOWN MULTI-AGENT THREADS WHILE NODE AND PYTHON SDKS PATCH SECURITY GAPS

Codex made parent-controlled sub-agent threads read-only in the TUI to prevent rejected server operations, while the Node and Python SDKs are closing security holes across GitHub Actions permissions, error handling, and dependency chains.

ai 106 shipped 2-min read

Elsewhere on the wire

@cachyos about 4 hours ago

PROTON 11.0 SHIPS WITH LOW-LATENCY VKD3D, MIRROR NETWORK EXPANDS

Proton 11.0 lands experimental low-latency graphics support while CachyOS adds a Hungary mirror to its distribution tier.

infra 45 shipped 1-min read

Elixir & Phoenix about 4 hours ago

ELIXIR TOOLING FIXES LAND AHEAD OF SYSTEM COMPATIBILITY PUSH

The compiler stops misinterpreting profile flags, string reversal handles invalid UTF8 correctly, and Ecto gains fragment column selection in a wave of fixes that shores up core library reliability.

elixir 11 shipped 1-min read

@dotnet about 4 hours ago

COMPLEX NUMBERS NOW IEEE 754 COMPLIANT, SKILLS EVALS GO VALLY-ONLY, UNSAFE EVOLVES

Complex<T> just shipped conformance to C23 Annex G special values, dotnet/skills completed its migration to the Vally eval harness, and the language is tightening unsafe semantics ahead of C# 13.

dotnet 141 shipped 2-min read

@spring-projects about 4 hours ago

SPRING BOOT TIGHTENS NULLABILITY CHECKS, FRAMEWORK CLEANS HOUSE

Spring Boot merged improvements to Nullaway 0.13.8 that correctly handle Map#remove nullability, while Spring Framework shipped targeted fixes for logging clarity and documentation.