Index
Features
- Self-testing and verificationSpaceXAI
On longer trajectories, checks its own work before moving on.
- Interactive and visual workSpaceXAI
Stronger first passes on visual and interactive projects — structure and visual language in one pass.
- Plan modeSpaceXAI
The agent explores and drafts a plan before edits. Only the session plan file may be edited until you approve; bash is not gated. Auto and always-approve do not skip plan review. Enter with /plan or Shift+Tab; the agent can enter plan mode on its own. Reopen with /view-plan.
- SubagentsSpaceXAI
Independent child sessions with their own context; they return a summary to the parent. Built-in types: general-purpose, explore (read/list/search only), and plan (no shell, no edits). Add types under .grok/agents/ or ~/.grok/agents/. Personas are behavioral overlays via /personas.
- SkillsSpaceXAI
Reusable folders with markdown, scripts, and resources (SKILL.md). Discovered from ./.grok/skills/, ~/.grok/skills/, plugin skills/, and [skills] paths. User-invocable skills appear as /<skill-name>. Capture a session as a skill with /skillify.
- Claude Code compatibilitySpaceXAI
Reads Claude Code marketplaces, plugins, skills, MCPs, agents, hooks, and instruction files (CLAUDE.md, Claude.md, CLAUDE.local.md, .claude/rules/) alongside .grok/ with no extra setup.
- HooksSpaceXAI
JSON hooks (command or HTTP) on lifecycle events. PreToolUse is the only blocking event (explicit deny, else fail-open). Personal ~/.grok/hooks/*.json; project .grok/hooks/*.json needs /hooks-trust or --trust. Also reads Claude Code and Cursor hook files.
- AGENTS.mdSpaceXAI
Loads AGENTS.md, Agents.md, AGENT.md, CLAUDE.md, and *.md in .grok/rules/ (plus .claude/rules/ and .cursor/rules/) from repo root down to cwd; deeper files win. --rules appends session-only text. grok inspect lists loaded files.
- Grok 4.6 in CursorCursor
Flagship model available in Cursor on Aug 12, 2026 with a one-week 2× usage boost.
9 of 73