Skills DirectorySkills Directory
SkillsLearnSecurityCategoriesDocsCommunityBlog
Sign InSubmit Skill
Skills Directory

Security-tested agent skills for Claude, coding agents, and AI workflows.

Directory

  • Browse Skills
  • All Skills A–Z
  • Claude Skills
  • Claude Code Skills
  • Agent Skills
  • Categories
  • Submit a Skill

Learn

  • Learn Hub
  • Install Claude Skills
  • Write SKILL.md
  • Skills vs MCP
  • Directories Compared

Security

  • Security
  • Methodology
  • Secure Claude Skills
  • Security Badges

Company

  • About
  • Community
  • Blog
  • API Docs
  • Advertise

2026 Skills Directory. All rights reserved.

Back to skills

Grok Lane

ASecurity

Orchestrate named, messageable local or remote Grok Build worker lanes from Claude Code — start detached work, collect a final answer, send follow-ups, resume the exact conversation, interrupt it, and archive it. Use when asked to delegate work to Grok, compare a Grok result, or use a named federated host.

2 stars
0 votes
0 copies
0 views
Added 9/22/2026
ai-agentsbash

Works with

claude codeterminalmcp

Security Analysis

A100/100

Scanned 9/22/2026

Install to Claude Code

$npx -y skills add antst/sessionbus-peers --skill grok-lane --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Grok Lane?

Add the live security badge to your README — it updates automatically with every re-scan.

Security grade badge for Grok Lane
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/antst-grok-lane/badge)](https://www.skillsdirectory.com/skills/antst-grok-lane)

More formats (shields.io, HTML) on the badges page.

Download Zip
Files
SKILL.md
---
name: grok-lane
description: Orchestrate named, messageable local or remote Grok Build worker lanes from Claude Code — start detached work, collect a final answer, send follow-ups, resume the exact conversation, interrupt it, and archive it. Use when asked to delegate work to Grok, compare a Grok result, or use a named federated host.
---

# Orchestrate Grok lanes

Use the process-attested `sessionbus.lane` MCP tool with `product: "grok"`.
It runs a headless Grok Build conversation as a daemon-owned durable Sessionbus peer. Do not execute `grok-peer-lane` through Bash.

## Remote host

For another host, use the same structured calls with `host` set. First call
`doctor` with arguments `["--json"]` and `list` with `["--all"]`.

Require capability `grok-lane` and doctor contract 2. Federation carries this Claude parent’s
attested context and returns terminal notices through grouped routing; never pass `--persistent`
or `--no-auto-archive` remotely. Use a remote absolute `-C` when cwd
matters. Hub disconnect is a hard failure; never fall back to SSH or local execution.

## Preflight

Call `doctor` with arguments `["--json"]`. Require contract version 2,
`runtime_ready: true`, and a ready native Grok adapter.
The nested `supervisor_reachable` field is diagnostic only because the Grok manager directly owns
ACP. Check `list --all` output plus current Sessionbus discovery, and retain the exact lane ID.
Messaging names are group-scoped; bare
lifecycle names can still be ambiguous in host-local lane state.

Headless Grok cannot approve prompts. Every lane is explicit `always-approve`
(`bypassPermissions`). Start it only when the user's task authorizes autonomous execution in the
selected cwd.

## Start, collect, and message

Pass the briefing as MCP `input`, never in `arguments`:

```json
{
  "product": "grok",
  "command": "start",
  "arguments": ["--name", "review-a", "--auto-archive-after", "300", "-"],
  "input": "<briefing text>"
}
```

`start` returns at `lane.ready`, not at the answer. Use one collector. A durable
`GROK_LANE_TERMINAL` pointer tells the owner when to collect; it is not the answer. A wait timeout exits 124 but
does not interrupt the turn. From the collected JSONL, select the `agent_message` final answer with
the same `turn_id` as the last `turn.completed`, and report `outcome` and `exit`.

The lane is a grouped peer. Use the `sessionbus` skill and
`sessionbus.send_message`; do not fall back to Claude's native messaging or
address a synthetic Grok row in Claude's registry. Then collect the resulting
serialized turn with `wait`. A message during active work is durably queued for
the next turn; it never creates a competing ACP driver.

For an explicit follow-up, call `resume` with arguments
`["review-a", "--timeout", "600", "-"]` and the follow-up as `input`.

Collect all debt first. Resume also unarchives and retains the exact UUID, transcript, cwd, name,
model, and reasoning policy.

## Retire safely

Call `status`, `interrupt`, and finally `archive` through the same lane tool,
passing `["review-a"]` as their arguments.

Default lanes are owned by this corroborated Claude process and archive when it exits. Use
`--persistent` only when the lane must survive its owner. It disables parent-exit cleanup only;
the normal 60-second terminal auto-archive grace still applies. Pair it with
`--no-auto-archive` only for indefinite idle retention, then archive explicitly.

Every child retains this immediate parent’s private group. Add `--inherit-groups` only when this
parent intentionally propagates its other groups; repeat `--group NAME` for child-specific groups.

Read [references/contract.md](references/contract.md) for the complete contract and
[references/install.md](references/install.md) when preflight reports a missing or incompatible
runtime.

Attribution

antstantst
View sourceMore from antst →
SSkills DirectorySkills Directory

Ship a skill? Prove it's safe.

Free 120-pattern security scan, letter grade, and an embeddable README badge.

Submit a skill

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments (0)

No comments yet. Be the first to comment!

SSkills DirectorySkills Directory

Ship a skill? Prove it's safe.

Free 120-pattern security scan, letter grade, and an embeddable README badge.

Submit a skill

Related Skills

Caveman

Ultra-compressed communication mode that cuts output tokens while keeping technical accuracy. Levels: lite, full, ultra and the wenyan variants. Use for /caveman, "caveman mode", "talk like caveman", "be brief" or "less tokens".

1066601 votes

Hyperplan

Adversarial multi-agent planning skill. Self-orchestrates 5 hostile category members (unspecified-low, unspecified-high, deep, ultrabrain, artistry) via team-mode for ruthless cross-critique debate, distills only the defensible insights, then MANDATORILY hands the distilled insight bundle to the `plan` agent for executable plan formalization. Use when planning needs maximum rigor and surfacing of weak assumptions, blind spots, and over-engineering. Triggers: 'hyperplan', 'hpp', '/hyperplan', ...

686011 votes

Mcp Code Execution

Routes multi-tool workflows through MCP servers for large datasets and pipelines. Use when Bash tool overhead is limiting throughput on data-heavy tasks.

3351 votes

catchup

Recovers the conversation and failed tool calls of a previous Codex, Claude Code, Antigravity, Cline, Copilot CLI, Cursor, DeepSeek Harness, Kimi, OpenCode, Pi Agent, or ZCode session. Use when the user says "catch up", "what did the last session do", "get me up to speed", "I switched agents", asks to recover/summarize a previous session before continuing, or asks to diagnose or report a catchup failure. Do NOT use for the current conversation, git history, or any non-agent log.

651 votes

math-skill

A comprehensive mathematical reasoning skill for AI assistants — handles arithmetic to research-level problems with rigorous step-by-step reasoning, systematic verification, and transparent uncertainty handling

381 votes
View all in ai-agents →