grok-cli-runtime

grok-cli-runtime is a skill for Claude Code, Codex from faeton/claude-grok-plugin. It costs 19 tokens per session (644 once invoked), scanned A, original, MIT.

A read-only helper for sending questions and code reviews to the Grok companion runtime from Claude Code.

In plain words
What is it for?
It helps a Claude Code subagent request advice, review code, or run an adversarial review while returning the companion's response unchanged.
Why use it?
It avoids manually constructing Grok commands and keeps reviews inside a safety wrapper that blocks changes and other risky actions.

Skill for Claude CodeCodex

Installs and runs on its own, but its text points at files inside its plugin — anything it tells you to read at a ${CLAUDE_PLUGIN_ROOT} path is only there once the plugin is installed. Installing the plugin gets both.

Part of the grok plugin — 3 skills, 8 commands, 1 agent 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/faeton/claude-grok-plugin/grok-cli-runtime
Any agent
npx skills add faeton/claude-grok-plugin --skill grok-cli-runtime
Clone the repo
git clone --depth 1 https://github.com/faeton/claude-grok-plugin

Made for: Claude Code, Codex.

Or install grok, the plugin that ships this one along with the rest of its 3 skills, 8 commands, 1 agent.

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-runtime

README.md
[![agentmods](https://agentmods.dev/badge/skills/faeton/claude-grok-plugin/grok-cli-runtime.svg)](https://agentmods.dev/skills/faeton/claude-grok-plugin/grok-cli-runtime)
Your own site
<a href="https://agentmods.dev/skills/faeton/claude-grok-plugin/grok-cli-runtime"><img src="https://agentmods.dev/badge/skills/faeton/claude-grok-plugin/grok-cli-runtime.svg" alt="Measured on agentmods" height="20"></a>
Per session 19 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 644 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 $0.00019 $0.00644
Opus 5 $0.00010 $0.00322
Sonnet 5 $0.00004 $0.00129
Haiku 4.5 $0.00002 $0.00064

Measured 3d ago against content hash c9019f06dc50, 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-runtime 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.

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.

plugins/grok/skills/grok-cli-runtime/SKILL.md · 32 lines

What it actually says

Grok Runtime

Use this skill only inside the grok:grok-consult subagent (and when reasoning about how grok commands shell out).

Primary helpers (one Bash call each, never combined):

  • node "${CLAUDE_PLUGIN_ROOT}/scripts/grok-companion.mjs" consult [flags] -- <question>
  • node "${CLAUDE_PLUGIN_ROOT}/scripts/grok-companion.mjs" review [flags]
  • node "${CLAUDE_PLUGIN_ROOT}/scripts/grok-companion.mjs" adversarial-review [flags] [focus text]

Execution rules:

  • The subagent is a forwarder, not an orchestrator. Invoke exactly once and return stdout unchanged.
  • Prefer the helper over hand-rolled git, raw grok strings, or any other Bash activity.
  • Always read-only. The companion already removes and denies Write/Edit/MultiEdit/NotebookEdit/Bash, disables subagents and memory, and runs grok under a watchdog. Never add flags that would re-enable editing, and never call grok directly — that bypasses the safety wrapper.
  • Leave --effort / --model unset unless the user explicitly asked. The companion auto-drops effort flags for models that do not support them.
  • Always run the companion in the foreground — never pass the companion's --background. Strip --wait/--background from the natural-language task text. Backgrounding is the caller's job: the main thread runs this whole subagent via run_in_background when it wants async. The companion's --background detaches its own worker, which gets orphaned the moment this subagent returns and its process tree is torn down (the worker then reconciles to failed / "Worker process exited without finishing"). Foreground is the only safe mode inside the subagent.

Command selection:

  • Free-form question or "what do you think of X" → consult. Put the whole question after --.
  • "Review my changes / this diff / this branch" → review (add --scope / --base if the user specified).
  • "Review, and specifically hunt for " → adversarial-review <X>.
  • Background jobs (created only by the direct /grok:consult / /grok:review slash commands, never by this subagent) are tracked: the user reads them back with /grok:status, /grok:result <id>, /grok:cancel <id>. Do not poll these yourself from the subagent.

Failure handling:

  • If the companion says grok is missing/unauthenticated, return that verbatim and stop (point the user at /grok:setup).
  • If it reports a timeout or stopReason cancellation, return that verbatim. Do not retry silently or answer yourself.
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 · 32 lines · 19 tokens per session scan A c9019f06dc50

Subscribe to this mod's changes

grok-cli-runtime is a skill published in the GitHub repository faeton/claude-grok-plugin (1 stars, last pushed 19d ago), licensed MIT. It adds 19 tokens to every session and 644 once invoked, about $0.0001 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.

Related

Other skills, from other repositories

grok-prompting

Internal guidance for composing Grok Build prompts for coding, review, diagnosis, and research tasks inside the Grok Claude Code plugin.

davit-khaburdzania/grok-plugin-cc · 31 tokens

grok-cli-runtime

Internal helper contract for calling the grok-companion runtime from Claude Code.

davit-khaburdzania/grok-plugin-cc · 19 tokens

grok-result-handling

Internal guidance for presenting Grok helper output back to the user.

davit-khaburdzania/grok-plugin-cc · 18 tokens

local-council-execution

Runs a local council when no external AI providers are configured. Spawns N independent Claude subagents (one per role, blind to each other) that each answer the question from a single assigned lens, then synthesizes their perspectives. Invoked by the ask command via --local, or when the user accepts the local-council…

hex/claude-council · 98 tokens

deep-execution

Executes agent-enhanced council queries by spawning parallel Claude subagents that each query a provider, evaluate response quality, ask follow-up questions, and return structured insights with confidence ratings and blind spot analysis. Invoked when the --agents flag is used or when complex architectural decisions…

hex/claude-council · 61 tokens

council-execution

Executes council queries by running the query pipeline across selected AI providers (Gemini, OpenAI, Grok, Perplexity), displaying formatted responses verbatim, and generating a synthesis of consensus, divergence, and recommendations. Invoked by the ask command during standard (non-agent) council queries.

hex/claude-council · 66 tokens