grok-cli-agent

grok-cli-agent is a skill for Claude Code, Codex from ImL1s/agent-cli-skills. It costs 88 tokens per session (607 once invoked), scanned A, original, MIT.

A command-line wrapper for running xAI’s Grok Build coding agent without an interactive terminal. It supports planning, worktrees, structured output, and subagents.

In plain words
What is it for?
Use it to ask Grok to inspect or change a repository, run plan-only reviews, start isolated worktrees, or fan work out to subagents.
Why use it?
It makes Grok work easier to run from scripts or an orchestrator, with options for isolated changes and parallel agent work.

Skill for Claude CodeCodex

Part of the agent-cli-skills plugin — 7 skills shipped together

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add skills/iml1s/agent-cli-skills/grok-cli-agent
Any agent
npx skills add ImL1s/agent-cli-skills --skill grok-cli-agent
Clone the repo
git clone --depth 1 https://github.com/ImL1s/agent-cli-skills

Made for: Claude Code, Codex.

Or install agent-cli-skills, the plugin that ships this one along with the rest of its 7 skills.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for grok-cli-agent

README.md
[![agentmods](https://agentmods.dev/badge/skills/iml1s/agent-cli-skills/grok-cli-agent.svg)](https://agentmods.dev/skills/iml1s/agent-cli-skills/grok-cli-agent)
Your own site
<a href="https://agentmods.dev/skills/iml1s/agent-cli-skills/grok-cli-agent"><img src="https://agentmods.dev/badge/skills/iml1s/agent-cli-skills/grok-cli-agent.svg" alt="Measured on agentmods" height="20"></a>
Per session 88 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 607 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00088 $0.00607
Opus 5 $0.00044 $0.00303
Sonnet 5 $0.00018 $0.00121
Haiku 4.5 $0.00009 $0.00061

Measured 3d ago against content hash 41c83d03bb8c, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

grok-cli-agent scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 3d ago.

The scan reads SKILL.md. This mod also ships 4 executable files (grok-exec.sh, lib/common.sh, lib/parse_stream_json.py, …), listed below but not scanned — reading those needs a real analyzer, not pattern matching.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

skills/grok-cli-agent/SKILL.md · 49 lines

What it actually says

Grok Build CLI (grok) Executor

When to use

  • User wants grok / Grok Build as the coding agent
  • Need Grok spawn_subagent, --agents JSON, or --worktree isolation from an orchestrator

Wrapper

./grok-exec.sh -C /path/to/repo -t 1800s "Implement X. NO git."
./grok-exec.sh -r -C /path/to/repo "Review lib/foo for races"
./grok-exec.sh -T -C /path/to/repo "Parallel research + implement"
./grok-exec.sh -f /tmp/bundle.txt -m grok-4.5 -e high -o json
./grok-exec.sh -w feat-branch -C /path/to/repo "Isolated worktree edit"

Flags: -m · -t · -C · -e effort · -o plain|json|streaming-json · -f · -r (plan + deny write/shell) · -s PROFILE · -T native subagent preamble · -w [NAME] worktree · --agents JSON · --max-turns · --json-schema · --rules · --disable-web-search · --no-git · -- passthrough.

Default model: grok-4.5 (grok models). No PTY needed — headless -p works on pipes. -C is absolutized before use.

Native multi-agent

Grok has first-class spawn_subagent (types: general-purpose, explore, plan), --agents JSON definitions, --no-subagents, --worktree isolation. -T asks the model to fan out freely and synthesize.

Discipline

  • Exit code untrusted — verify results yourself
  • Orchestrator owns git (--no-git)
  • Long prompts: always -f file

Raw surface

See grok --help: permission modes, sandbox, json-schema, tools allow/deny, resume/continue, grok agent (stdio/headless/serve/leader), worktree, export, doctor.

Files

What ships with it

4 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 3d ago First seen · 49 lines · 88 tokens per session scan A 41c83d03bb8c

Subscribe to this mod's changes

grok-cli-agent is a skill published in the GitHub repository ImL1s/agent-cli-skills (2 stars, last pushed 1mo ago), licensed MIT. It adds 88 tokens to every session and 607 once invoked, about $0.0004 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.