Give it a reproducible task
Supply the failing behavior, relevant files and a test command. Ask the agent to explain the cause before making a focused change.

SpaceXSI (formerly xAI) · Coding tools
SpaceXSI’s coding agent, with an open-source terminal interface that can run against local models, plus web and mobile versions and memory that carries across sessions.
Part of Grok. See its other features, models and news on the Grok page.
Delegating a bounded repository change through a terminal coding workflow.
Available on every plan, drawing on the shared Grok and Cursor usage pool. Use a recoverable working copy, scope file access and inspect the patch; a claim that tests passed should come with the commands and results.
Give it one reproducible bug and a test command. Review the diff and rerun the test.
Supply the failing behavior, relevant files and a test command. Ask the agent to explain the cause before making a focused change.
Run a check that would fail before the fix. Inspect dependency changes and unrelated edits, then try the changed behavior in the application.
No stories mention Grok Build in the current feeds. Browse all AI news →

Anthropic
Compare a repository task, the patch and the tests actually run.

OpenAI
Try the same coding task with another coding agent.

Cursor (SpaceX)
Consider an editor-centered workflow instead of a terminal-first one.