← All Grok Build issues

This week in Grok Build · Aug 10, 2026

Grok Build August 10: xAI Ships Grok Build 1.0.0, the Terminal Coding Agent Leaves Beta Free Under Apache 2.0 With grok-build-0.1 and Grok 4.5 Unchanged

Get the next issue in your inbox. Weekly · Free

unsubscribe anytime.

TLDR;

Grok Build 1.0.0 landed on August 7, the first stable release of xAI's terminal coding agent, eleven weeks after the CLI first appeared in early beta. It is a stability and interface release, not a model release: the model served through the xAI API is still grok-build-0.1 and the default agent in the terminal is still Grok 4.5. Dashboard rows now summarize what the agent did in the previous turn, auto theme detection works over SSH and inside tmux, markdown tables reflow inside cells on narrow panes, and typing exit or quit in the dashboard exits the CLI. Grok Build stays free to install under the Apache 2.0 license with a single shell command, and Elon Musk confirmed within hours that a version aimed at non-technical users is already in progress with no timeline. Upgrade to 1.0.0 and re-test your MCP servers this week.

Grok Build CLI

Grok Build 1.0.0 is out (latest, August 7): xAI's terminal coding agent leaves beta with turn summaries in the dashboard, SSH and tmux theme detection, and reflowing markdown tables

  • August 7: xAI shipped Grok Build 1.0.0, the first stable release of its terminal coding agent.
  • It is stability and interface polish, not a new model: the xAI API still serves grok-build-0.1 and the terminal default is still Grok 4.5.
  • The Extensions modal now groups items alphabetically with collapsible Skills sections, auto theme detection works over SSH and inside tmux, and markdown tables reflow on narrow panes.
  • It stays free under Apache 2.0 with a single-command install, and Elon Musk confirmed a version for non-technical users is already in progress.

Editor’s read: A 1.0 that ships no model change is xAI telling you the CLI is now safe to standardize on. The MCP tool image handling and permission-prompt work are the two areas to re-test first, since both changed underneath you.

Developer Tools & Community

Open-source ADLC team-skills brings team coding standards to any agent supporting the Agent Skills standard

  • 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 Claude Code, Codex, 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: Relevant to Grok Build 1.0.0 specifically because the Extensions modal now groups Skills into collapsible sections, so a large shared skills repo is finally navigable in the dashboard.

Community & Independent Developers

Hoplite (YC S26) launches cloud coding agent deployment with a custom in-house harness 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.
  • The founders built a custom agent harness in-house rather than wrapping Codex or Claude Code, for independence from vendor release cycles, and the launch offer is $100 in free credits with the code HACKERNEWS.

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.

Competitor Dev Tools

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

  • 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.

Editor’s read: Muse Code arrived in beta two days before Grok Build reached 1.0. If you are choosing a terminal agent this month, Grok Build's Apache 2.0 license and single-command install are the concrete advantages, not the model.

Research & Safety

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: The worm harvests xAI-adjacent secrets too, since it sweeps roughly 200 glob patterns for sensitive files. Rotate any API key that lives in a dotfile, not just the ones the post names.

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 at a cost of $2,600.

Editor’s read: No Grok model appears on the leaderboard yet, which is the gap to watch now that the CLI has shipped 1.0 while the served model stays at grok-build-0.1.

Action Items

Immediate

  • Upgrade to Grok Build 1.0.0 with the single-command install, and re-test permission prompts and MCP tool image handling, both of which changed in this release.
  • Re-check auto theme detection if you run Grok Build over SSH or inside tmux, since detection now works in both.
  • Audit your lockfiles for keyv, flat-cache, file-entry-cache and cacheable-request, and rotate npm tokens, GitHub PATs, AWS credentials, Kubernetes secrets, Vault tokens and Stripe and Slack keys if you installed during the August 4 window.
  • Confirm whether you are still on grok-build-0.1 through the xAI API or on Grok 4.5 in the terminal, because 1.0.0 changed neither.

By August 17

  • Reorganize your Skills into groups now that the Extensions modal sorts alphabetically with collapsible Skills sections.
  • Compare Grok Build 1.0.0 against Meta's new Muse Code beta on cost and repository scale before committing a team to either.
  • Pin your Grok Build version in team setup docs, since 1.0.0 is the first release stable enough to standardize on.

All Resources

Every Monday morning

This week in Grok Build

Five minutes, every Monday - the tools, releases and tactics for developers.

unsubscribe anytime.

This week in AI coding

Five minutes, every Monday - the tools, releases and tactics for developers.

unsubscribe anytime.