Who contributed to OpenAI on August 6, 2026?

4 developers shipped this update, including felixxia-oai, github-actions[bot], copyberry, and Hayden.

What were the notable OpenAI updates?

0.146.1, Move explicit skill selection into the skills crate (#37177), and Load host skill roots through the skills extension.

OpenAI

Codex, the SDKs, and the engine behind ChatGPT

Pick a date

Topics: AI / ML Full archive →

The Wire · Showcase

CODEX HARDENS SKILL SELECTION AND SESSION LOOKUP, NODE SDK CLOSES LINTER GAPS

By RepoJournal · Filed 06:58 UTC on August 6, 2026 · About OpenAI

4 people shipped this

Hayden @HAYDEN-OAI 5 cited

copyberry @copyberry 3 cited

felixxia-oai @felixxia-oai 1 cited

github-actions[bot] @github-actions[bot] 1 cited

Codex decouples explicit skill selection into a dedicated crate while tightening TUI session validation, as the Node SDK completes its oxlint cleanup stack across five rule categories.

Codex shipped 0.146.1 [1] with safer automatic-review defaults for cyber-capable models, but the real work is structural. The team moved explicit skill selection into the skills crate [2], decoupling mention selection from core skill loading through a new `ExplicitSkillLookup` interface. In parallel, the TUI hardened named session lookup [4] by validating rollout identity, collection, source, and model provider before selecting local sessions, while a separate effort [3] consolidates host skill roots through the skills extension to preserve caching and namespacing. Remote environment testing [5] now covers provisioning readiness before selection, verifying that transport materialization and capability-root context are available when needed. Over on openai-node, Hayden closed the oxlint cleanup stack [6] [7] [8] [9] [10], removing five compatibility exceptions from the linter config: no-template-curly-in-string, no-useless-constructor, prefer-exponentiation-operator, unicorn/escape-case, and unicorn/no-zero-fractions. No source changes were required; all handwritten files already satisfied the rules.

Action items

→ Review and deploy Codex 0.146.1 to pick up cyber-model review defaults [ref:2] openai/codex [plan]
→ Verify openai-node linter baseline against latest oxlint rules to stay in sync openai/openai-node [monitor]
→ Test explicit skill selection decoupling in your environment [ref:1] openai/codex [monitor]

References

[1] 0.146.1 ↗ openai/codex
[2] Move explicit skill selection into the skills crate (#37177) openai/codex
[3] Load host skill roots through the skills extension ↗ openai/codex
[4] Harden named session lookup in the TUI ↗ openai/codex
[5] Test remote environments reported ready before selection ↗ openai/codex
[6] chore(lint): enable no-template-curly-in-string (#2071) openai/openai-node
[7] chore(lint): enable no-useless-constructor (#2072) openai/openai-node
[8] chore(lint): enable prefer-exponentiation-operator (#2073) openai/openai-node
[9] chore(lint): enable unicorn/escape-case (#2074) openai/openai-node
[10] chore(lint): enable unicorn/no-zero-fractions (#2075) openai/openai-node

Quick answers

What shipped in OpenAI on August 6, 2026?: Codex decouples explicit skill selection into a dedicated crate while tightening TUI session validation, as the Node SDK completes its oxlint cleanup stack across five rule categories. In total, 56 commits, 56 pull requests, and 6 releases landed.
Who contributed to OpenAI on August 6, 2026?: 4 developers shipped this update, including felixxia-oai, github-actions[bot], copyberry, and Hayden.
What were the notable OpenAI updates?: 0.146.1, Move explicit skill selection into the skills crate (#37177), and Load host skill roots through the skills extension.

CODEX HARDENS SECURITY, SDKS MODERNIZE TOOLCHAINS

Codex now prompts users before trusting local directories, closing a prompt injection exposure, while both Node and Python SDKs shift to Oxlint and Oxfmt tooling.

ai 114 shipped

OpenAI Aug 4

GPT-5.5 LANDS ACROSS NODE AND PYTHON SDKS, CODEX HARDENS MCP CONFORMANCE

OpenAI shipped gpt-5.5 model support and tool metadata across both JavaScript and Python SDKs while Codex locked down a full MCP client conformance regression suite to prevent protocol slippage.

ai 115 shipped

OpenAI Aug 3

CODEX HARDENS LOGIN FLOW WITH ONBOARDING HINTS AND METADATA PROTECTION

Three critical fixes shipped across auth, data persistence, and usage tracking to prevent thread metadata loss and secure OAuth callback handling.

ai 6 shipped

OpenAI Aug 2

CODEX SHIPS AGENT PLUGIN OVERHAUL AND TUI PERFORMANCE FIXES

Codex now handles portable Agent Plugins with schema-declared manifests and cuts terminal redraw overhead by caching screen dimensions instead of querying on every frame.

ai 12 shipped

Elsewhere on the wire

AI Agents Aug 6

LANGCHAIN HARDENS ANTHROPIC TOOL CALLING AND GIT CREDENTIALS AS OPENHANDS SHIPS 1.10.0

LangChain fixed a critical credential leak in evaluation runs while Anthropic tool schemas get proper fallback handling, as three major agent frameworks push releases across the ecosystem.

ai-agents 54 shipped

Local LLMs Aug 6

OLLAMA FIXES KV CACHE CORRUPTION, LLAMA.CPP HARDENS FILE SEARCH, VLLM KILLS MOE LEGACY CODE

Ollama patched a critical cache corruption bug that corrupts lazy snapshots during multi-request prefills, while llama.cpp hardened Windows junction traversal and vLLM removed 14 hardcoded MoE shards across quantization methods.