Updated August 26, 2026: aligned this guide with xAI's current Grok 4.6 Build page, feature list, official install command, and safer inspect-before-run guidance.
Verdict: Grok Build is xAI's terminal coding agent, currently presented as powered by Grok 4.6 and available to try for free. Its plan mode, skills, plugins, hooks, MCP servers, subagents, sandboxing, memory, and AGENTS.md support make it capable. That capability is exactly why installation and permissions deserve scrutiny.
What is Grok Build?
Grok Build brings an xAI coding agent into the terminal. It can inspect a repository, propose plans, edit files, run commands, use tools, and work with persistent project guidance. xAI positions it for complex coding work rather than as a simple autocomplete utility.
The current xAI product page says Grok Build is powered by Grok 4.6. Any article that still treats the launch model as current should be updated and dated because model assignments can change independently of the CLI itself.
Official installation command
xAI currently publishes this command:
curl -fsSL https://x.ai/cli/install.sh | bash
It downloads a shell script and immediately pipes it to Bash. That is convenient, but it gives the remote script the permissions of your shell before you inspect it.
A safer inspect-before-run installation
For a security-conscious workstation, separate download from execution:
curl -fsSL https://x.ai/cli/install.sh -o /tmp/grok-build-install.sh
sed -n '1,240p' /tmp/grok-build-install.sh
shasum -a 256 /tmp/grok-build-install.sh
bash /tmp/grok-build-install.sh
Review where the script downloads binaries, which directories it changes, whether it asks for elevated privileges, and how it updates your shell path. The checksum records what you reviewed; it does not prove the file is safe unless xAI publishes a trusted checksum for comparison.
Only run the final Bash command after the review. In a managed environment, use the organization's package approval and endpoint-security process.
What does x.ai/cli/install.sh install?
The exact script can change, so the correct answer is the content of the script at the time you run it. Do not rely on a static article to enumerate every file forever. Download the script, inspect its destinations and network calls, and compare the installed binary with xAI's current documentation or release information.
This page intentionally avoids pretending that a changing remote script has a timeless file list.
Is there an official PowerShell installer?
xAI's public Grok Build page displayed a Bash install command during the August 26, 2026 check. We did not find an official PowerShell command on that page. Windows users should follow xAI's current documentation for Windows or WSL rather than translating the shell script into an unverified one-liner.
If the docs do not support your environment, stop there. A community workaround can be useful, but it should be labelled and reviewed separately from the official install path.
Grok Build features
| Feature | What it does | Safety implication |
|---|---|---|
| Plan mode | Produces a reviewable plan before edits | Keep edits blocked until approval |
| Skills | Loads reusable task instructions | Review skill sources and scope |
| Plugins | Bundles skills, agents, hooks, and MCP servers | Installation expands the tool surface |
| Hooks | Runs code around lifecycle events | Treat hook code as executable software |
| MCP servers | Connects external tools and data | Limit credentials and permissions |
| Subagents | Splits work across agents | Each agent must inherit safe boundaries |
| Sandbox | Restricts commands and file access | Configure and test before unattended use |
| AGENTS.md | Supplies repository guidance | Keep it version-controlled and reviewable |
| Headless operation | Enables scripts and CI | Add budgets, logs, approvals, and timeouts |
xAI says skills can be auto-invoked when a task matches, and that AGENTS.md, plugins, hooks, and MCP servers work with the product. Automatic capability discovery is convenient, but it means a repository can affect agent behaviour. Review repository instructions and plugin manifests like code.
Plan mode and approvals
xAI's product page says plan mode blocks every edit until the user approves the plan. That is a valuable boundary for unfamiliar repositories and high-impact changes. It is not a complete security model: an approved plan can still contain an unsafe command, broad file scope, or unnecessary dependency.
Review the actual diff and command list, not only the plan's summary. Keep production deployment, secrets, billing, external messages, and destructive actions behind separate explicit approvals.
Sandboxing, headless use, and CI
Headless agents need tighter controls than interactive agents because nobody is watching each step. Use a disposable worktree or container, provide read-only credentials where possible, set a command allowlist, cap runtime and spend, and export logs and diffs as artifacts.
A safe CI progression is:
- Read-only analysis.
- Suggested patch with no push.
- Patch in an isolated branch with tests.
- Human-reviewed pull request.
- Deployment handled by the existing release system.
Do not give a new coding agent direct production credentials just because it supports headless mode.
Is Grok Build free?
xAI's page says Grok Build is available to try for free. That is not a guarantee of unlimited permanent usage. Included usage, model availability, account requirements, and API pricing can change. Check the current product and account pages before planning a sustained workload.
Grok Build versus a programmable SDK
Grok Build is terminal- and workflow-oriented. A programmable agent SDK such as the Cursor SDK is designed to be embedded in scripts and applications. Both can automate coding work, but the integration surface, model provider, runtime, permissions, and billing differ.
Choose based on the control plane you need, not the model name alone.
Compatible-workflow notes and related reading retained from the verified current article
Works with what you already use
A recurring failure mode for AI dev tools is forcing engineers to abandon their existing workflows. Grok Build’s team appears to have learned from that mistake. Out of the box, it picks up:
AGENTS.md— a project-level instructions file that the agent reads before it does anything, similar to the convention popularized by other agent CLIs.- Plugins, hooks, and skills — capabilities packaged so they can be shared across a team or installed from a marketplace.
- MCP servers — Model Context Protocol integrations that let Grok Build talk to external tools and data sources without bespoke glue code.
Drop the CLI into a repo that already has these conventions and it picks them up automatically. The demo on the Grok Build site shows the agent installing a browser-review plugin from a community marketplace, hooking it into the current session, and exposing its skills — all without leaving the terminal. The marketplace concept is interesting on its own: teams can publish and share capabilities the way they share packages, instead of every engineer reinventing the same wrapper around the same internal API.
Compare
Related Models
Recent Launches
- Elon’s xAI Unleashes Grok 3 for Free—Catch It Before the Servers Sizzle and Pop!
- Grok now buffed up: Elon Musk announced Grok is now much faster and now cites legacy media as resources.
- Grok Spreadsheet Editing and Studio Features Revealed
Latest News
- Grok 4 Benchmarks Explained: Why Its Performance is a Game-Changer
- Grok by X: How iOS Users in the US Can Tap Into Powerful AI Insights
Continue Reading
FAQ
What is x.ai/cli/install.sh?
It is the remote shell script used by xAI's official Grok Build install command. Download and inspect it before execution.
How do I install Grok Build safely?
Save the script to a temporary file, inspect it, record a checksum, and run it only after the destinations and commands are acceptable.
Is there an official PowerShell installer?
The public product page showed Bash, not PowerShell, during this check. Use current xAI Windows or WSL documentation instead of an invented translation.
Is Grok Build free?
xAI says it is available to try for free. Ongoing limits and pricing can change.
Does Grok Build support AGENTS.md and sandboxing?
Yes. xAI's product page lists AGENTS.md support and a broader feature set that includes sandbox-oriented controls.
Official sources
The Kingy Brief
Get the next Kingy Brief.
Source-checked AI changes, original tests and one practical thing to try.
Free · Choose your subjects · Double opt-in · Unsubscribe anytime