Who contributed to OpenAI on August 24, 2026?

3 developers shipped this update, including Eric Traut, pakrym-oai, and copyberry.

What were the notable OpenAI updates?

Remove the Plan mode composer nudge (#40200), Remove the Plan mode composer nudge, and Preserve content metadata when truncating messages (#40264).

OpenAI

Codex, the SDKs, and the engine behind ChatGPT

Pick a date

Topics: AI / ML Full archive →

The Wire · Showcase

Codex Drops the 'Plan' Nudge

By RepoJournal · Filed 06:57 UTC on August 24, 2026 · About OpenAI

3 people shipped this

copyberry @copyberry 3 cited

Eric Traut @etraut-openai 1 cited

pakrym-oai @pakrym-oai 1 cited

OpenAI rolled back the Plan mode prompt in Codex, quietly removing the 'Create a plan?' nudge from the composer and the friction it brought to drafting.

OpenAI stripped the Plan mode composer nudge out of Codex today [1], so drafts containing the word 'plan' no longer trigger that persistent prompt or its Escape-to-dismiss behavior. The same change landed twice [2], one from an engineer and one from a bot, which is a good reminder to check your merge queue for duplicates. More meaningful for state fidelity, message truncation now preserves passthrough content metadata and classifications as `AnnotatedContent` [3], so positions stay in sync after the token budget cuts in. Guardian review threads finally get their own thread source, `guardian_review`, instead of being lumped in with generic subagents [4], which cleans up analytics and rollout attribution. And Codex now treats `cua_repl` like `node_repl` for Guardian evidence gathering and computer-use policy [5], so REPL-backed tools behave consistently across the stack.

Action items

→ Pull the latest codex main to get the Plan mode nudge removal and metadata preservation fixes openai/codex [immediate]
→ Update any analytics dashboards that relied on the generic 'subagent' thread source for Guardian reviews openai/codex [plan]

References

[1] Remove the Plan mode composer nudge (#40200) openai/codex
[2] Remove the Plan mode composer nudge ↗ openai/codex
[3] Preserve content metadata when truncating messages (#40264) openai/codex
[4] Distinguish Guardian review threads from subagents ↗ openai/codex
[5] Support `cua_repl` as a Node REPL-backed MCP server ↗ openai/codex

Quick answers

What shipped in OpenAI on August 24, 2026?: OpenAI rolled back the Plan mode prompt in Codex, quietly removing the 'Create a plan?' nudge from the composer and the friction it brought to drafting. In total, 22 commits, 22 pull requests, and 2 releases landed.
Who contributed to OpenAI on August 24, 2026?: 3 developers shipped this update, including Eric Traut, pakrym-oai, and copyberry.
What were the notable OpenAI updates?: Remove the Plan mode composer nudge (#40200), Remove the Plan mode composer nudge, and Preserve content metadata when truncating messages (#40264).

CODEX TIGHTENS THREAD CLASSIFICATION, NODE SDK ADDS X.509 TRANSPORT

Codex now requires content classification for all thread sources and extension fragments, while the Node SDK introduces caller-attested X.509 transport for hardened auth flows.

ai 26 shipped

OpenAI Aug 22

OpenAI SDKs Tighten Security and Polling Internals

OpenAI shipped a wave of SDK refactors overnight, moving webhook signature checks and file polling into SDK-owned helpers while Codex expands browser and computer-use policy controls.

ai 121 shipped

OpenAI Aug 21

OPENAI NODE SHIPS FOUR SECURITY FIXES, STREAMING SAFETY AND CREDENTIAL LEAKS PATCHED

If you run the official OpenAI Node SDK, today's security patch train is the one you can't skip.

ai 167 shipped

OpenAI Aug 20

OPENAI-NODE SECURITY BLITZ PATCHES STREAMING EXPOSURES

Five security fixes landed in openai-node overnight, closing unauthenticated proxies, unbounded buffers, and path leaks that could have billed your account or leaked patient data.

ai 155 shipped

Elsewhere on the wire

AI Agents Aug 24

CLINE SHIPS V4.1.15 AS TELEMETRY BUGS SURFACE ACROSS THE AGENT STACK

Cline's long-awaited auto-approve fix landed overnight, but it's the telemetry drop in 4.1.11 that should have you checking your session metrics.

ai-agents 44 shipped

Local LLMs Aug 24

VLLM'S P2P WEIGHT SYNC CUTS RL MOE TRANSFER COSTS

Forget all-gathering experts during RL rollout: vLLM's new sharded_rdt backend lets each worker pull only the slice it needs, turning a total_bytes transfer into a total_bytes divided by num_workers one.