OpenAI Codex August 10: Codex CLI 0.147.0 Ships Portable Agent Plugins, the --approve-for-me Flag and Cursor Skill Import, and Sottiaux Says Today's Codex Will Seem Primitive in 2 to 3 Months

OpenAI Codex CLI 0.147.0 (latest, August 7) adds portable Agent Plugins, the --approve-for-me flag, Cursor skill import and the opt-in MCP 2026-07-28 protocol

• August 7: the release and the matching Codex changelog entry land together.
• Codex CLI 0.147.0 installs portable Agent Plugins and searches local, personal, workspace and remote catalogs, adds --approve-for-me for automatically reviewed approvals, imports Cursor-managed skills without creating duplicates, and supports the opt-in MCP 2026-07-28 protocol with paginated discovery and non-blocking server startup.
• It also redacts secrets and complete bearer tokens from displayed commands, and removes the deprecated codex exec --full-auto flag in favor of --sandbox workspace-write.

Editor’s read: The removal of codex exec --full-auto is the breaking change hiding in the Chores list. Grep your CI and shell scripts for it today and replace it with --sandbox workspace-write before the next upgrade.

OpenAI Codex CLI 0.146.1 applies safer automatic-review defaults for cyber-capable models

• August 5: a bug-fix release that applies safer automatic-review defaults for cyber-capable models and explains permission changes in the terminal interface, two days before --approve-for-me arrived in 0.147.0.

OpenAI Codex roadmap: core products lead Thibault Sottiaux says today's Codex will seem primitive in 2 to 3 months, with Codex now past 8 million users

• August 4: The New Stack reported on the Codex roadmap, quoting OpenAI core products lead Thibault Sottiaux saying Codex will seem primitive in 2 to 3 months.
• Codex has passed 8 million users since a new GPT-5 model launched for it in early July, OpenAI is working to move Codex beyond tasks limited to a developer's computer, and the planned purchase of Ona, which builds secure cloud development environments, is framed as the next phase, where agents keep working in a customer's cloud after the laptop that started the job is closed.

Editor’s read: Read this as a warning about where your workflow assumptions break. If you have wired Codex into local-only scripts and laptop state, the persistent-cloud direction will cost you a rewrite rather than an upgrade.

Open-source ADLC team-skills brings team coding standards to Codex and Claude Code as agent skills

• August 4: tikalk's adlc-team-skills repo trended on Hacker News with a set of agent skills implementing an Agentic Software Development Lifecycle, working with any agent that supports the Agent Skills standard, including Codex, Claude Code, OpenCode, Cursor and Copilot.
• team-boot injects the team's rules, personas and decisions at session start as a roughly 100-token index with full rules loaded on demand, mission-brief enforces a specification-first contract before implementation, and levelup extracts session learnings as Context Directive Records committed back to the team repo.

Editor’s read: Good timing with Codex CLI 0.147.0, which now imports Cursor-managed skills: a shared team-skills repo is becoming portable across every agent your team runs.

Hoplite (YC S26) launches cloud coding agent deployment with Codex subscription support and $100 in free credits

• August 3: Hoplite, at hoplite.sh, posted its Launch HN for deploying coding agents in the cloud with tooling aimed at QA-ing features rather than reviewing code, porting over your local sessions, memories and MCP servers.
• You can connect a Codex subscription to use OpenAI models through it, the launch offer is $100 in free credits with the code HACKERNEWS, and the founders deliberately built a custom agent harness in-house rather than wrapping Codex or Claude Code, for independence from OpenAI and Anthropic release cycles.

AI coding cognitive debt: one developer retypes every line of LLM-generated code by hand, and 530 points on Hacker News agreed it is worth arguing about

• August 3: Ankur Sethi's essay Prevent cognitive debt by manually retyping LLM-generated code hit the Hacker News front page with over 530 points.
• He argues that passively accepting AI output creates cognitive debt, the risk of no longer understanding your own codebase, and accepts roughly 2x efficiency instead of 10x because retyping each edit forces him to catch errors, refactor as he goes and build a spatial map of the project's architecture.

Editor’s read: Worth reading the same week --approve-for-me lands. The more approvals you automate, the more deliberately you have to read what shipped.

Meta launches Muse Code, a terminal coding agent for large code bases, against OpenAI Codex CLI and Claude Code

• August 5: Meta released Muse Code in beta, a terminal coding agent powered by its Muse Spark coding model, which Meta says can accomplish complete software engineering tasks across large repos including planning, writing and validating changes.
• It runs parallel sub-agents in isolated worktrees, and in Meta's testing it built six features for a game simultaneously with no collisions and without touching the user's working copy.
• Meta's AI chief named cost-effectiveness as the differentiator against Codex and Claude Code.

Editor’s read: Meta is competing on price, which is the one axis where Codex users notice immediately. Watch for published rates before you take the cost-effectiveness claim seriously.

npm supply chain attack: keyv, flat-cache and 434 more packages hit by the Shai-Hulud worm, over 2 billion monthly installs affected

• August 4: attackers compromised the GitHub account of the keyv maintainer, at roughly 127M weekly npm downloads.
• 11 packages were directly compromised plus 434 more through worm propagation, over 2 billion monthly installs in total, including keyv, flat-cache, file-entry-cache and cacheable-request.
• A preinstall script in setup.mjs and Math_Symbol.js runs during npm install, downloads the Bun runtime and harvests npm tokens, GitHub PATs, AWS credentials, Kubernetes secrets, Vault tokens and Stripe and Slack keys, then exfiltrates RSA-encrypted data to attacker-controlled GitHub repos.

Editor’s read: Codex CLI 0.147.0 hardened plugin isolation and secret redaction in the same week, which is the right direction but no help for credentials already on disk. Audit lockfiles and rotate tokens.

MirrorCode benchmark: Claude Opus 4.7 rebuilt a 16,000-line Go bioinformatics toolkit in 14 hours for $251

• August 3: Epoch AI and METR published MirrorCode, a benchmark asking what the largest software project AI can complete on its own is, requiring models to reimplement entire programs end to end without access to the original source across 25 targets.
• The headline result is Claude Opus 4.7 reimplementing gotree, a bioinformatics toolkit of roughly 16,000 lines of Go with more than 40 commands, in 14 hours at a cost of $251, against an estimate of 2 to 17 weeks for human engineers.
• One task ran 19 days without human intervention at a cost of $2,600.