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

Agent Lanes

ASecurity

Start, collect, message, resume, and archive durable Codex, Claude, Grok, or Qwen worker lanes from Grok. Use for delegation to any supported target product, local or remote parallel work, and follow-ups.

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

Works with

terminalmcp

Security Analysis

A100/100

Scanned 9/22/2026

Install to Claude Code

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

Installs into .claude/skills of the current project.

Are you the author of Agent Lanes?

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

Security grade badge for Agent Lanes
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/antst-agent-lanes/badge)](https://www.skillsdirectory.com/skills/antst-agent-lanes)

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

Download Zip
Files
SKILL.md
---
name: agent-lanes
description: Start, collect, message, resume, and archive durable Codex, Claude, Grok, or Qwen worker lanes from Grok. Use for delegation to any supported target product, local or remote parallel work, and follow-ups.
---

# Agent lanes

Use the attested `sessionbus.lane` MCP tool to delegate from Grok without attaching a second
driver to this Grok conversation. It executes the exact packaged runtime, retains this Grok peer as
the lifecycle and communication parent, and returns `exit`, `stdout`, and `stderr`. Set `product` to
`codex`, `claude`, `grok`, or `qwen`, put the lifecycle verb in `command`, pass native trailing arguments in
`arguments`, and pass briefings as `input`. Add `host` for federation. Do not shell-execute a lane
launcher when the MCP tool is available; every lifecycle example below is an `sessionbus.lane`
argument object.

Example:

```json
{"product":"claude","command":"start","arguments":["--name","claude-review","--permission-mode","dontAsk","-"],"input":"Review the requested change and report the result."}
```

## Preflight

```json
{"product":"codex","command":"doctor","arguments":["--json"]}
{"product":"claude","command":"doctor","arguments":["--json"]}
{"product":"grok","command":"doctor","arguments":["--json"]}
{"product":"qwen","command":"doctor","arguments":["--json"]}
{"product":"codex","command":"list","arguments":["--all"]}
{"product":"claude","command":"list","arguments":["--all"]}
{"product":"grok","command":"list","arguments":["--all"]}
{"product":"qwen","command":"list","arguments":["--all"]}
```

All four unified daemon lanes require contract version **2**;
Claude also requires `claude_logged_in: true`, while Grok requires `grok_available: true` and no
`grok_error`. Qwen requires `ready: true`, including ACP/archive/profile integration evidence.
Require each doctor to report its target-specific readiness before using
that product. Fail closed: if a required field is false or missing, do not start that product and
report the lane cell as blocked; never "try anyway." Do not apply one product's contract version to
the other.

For a remote host, add `"host":"HOST"` to the same tool call only after the host advertises the
matching capability. Never fall back to SSH. Remote lifecycle and message delivery require the hub.

Every child always joins its own private group and this Grok parent’s private group. Do not copy
other parent groups unless the launch deliberately includes `--inherit-groups`. Use repeatable
`--group NAME` for child-specific membership. `--no-inherit-groups` resets optional inheritance
without removing the parent anchor; omission on resume restores the durable choice.

## Start and collect

Never place a briefing on argv. Put it in `input`:

```json
{"product":"codex","command":"start","arguments":["--name","codex-review","-"],"input":"BRIEFING"}
{"product":"claude","command":"start","arguments":["--name","claude-review","--permission-mode","dontAsk","-"],"input":"BRIEFING"}
{"product":"grok","command":"start","arguments":["--name","grok-review","-"],"input":"BRIEFING"}
{"product":"qwen","command":"start","arguments":["--name","qwen-review","--no-yolo","-"],"input":"BRIEFING"}
```

`lane.ready` means the worker is addressable; it is not the answer. Collect each lane with one
consumer:

```json
{"product":"codex","command":"wait","arguments":["codex-review","--timeout","300"]}
{"product":"claude","command":"wait","arguments":["claude-review","--timeout","300"]}
{"product":"grok","command":"wait","arguments":["grok-review","--timeout","300"]}
{"product":"qwen","command":"wait","arguments":["qwen-review","--timeout","300"]}
```

Take the last `turn.completed`, then the `agent_message` final answer with the same `turn_id`.
Report outcome and exit. A wait timeout exits 124 without interrupting the worker.

## Message, follow up, and retire

Use the installed `sessionbus` tools to send ordinary messages to a lane's current peer name or
session identity. A terminal pointer still requires `wait`; do not answer it conversationally.

```json
{"product":"codex","command":"resume","arguments":["codex-review","-"],"input":"FOLLOW-UP"}
{"product":"claude","command":"resume","arguments":["claude-review","-"],"input":"FOLLOW-UP"}
{"product":"grok","command":"resume","arguments":["grok-review","-"],"input":"FOLLOW-UP"}
{"product":"qwen","command":"resume","arguments":["qwen-review","-"],"input":"FOLLOW-UP"}
{"product":"codex","command":"interrupt","arguments":["codex-review"]}
{"product":"claude","command":"interrupt","arguments":["claude-review"]}
{"product":"grok","command":"interrupt","arguments":["grok-review"]}
{"product":"qwen","command":"interrupt","arguments":["qwen-review"]}
{"product":"codex","command":"archive","arguments":["codex-review"]}
{"product":"claude","command":"archive","arguments":["claude-review"]}
{"product":"grok","command":"archive","arguments":["grok-review"]}
{"product":"qwen","command":"archive","arguments":["qwen-review"]}
```

Collect outstanding debt before resume. Resume preserves the same transcript. Default lanes belong
to this corroborated Grok peer and archive when it exits; use `--persistent` only when survival is
required. Persistence disables parent-exit cleanup only; the normal 60-second terminal auto-archive
grace still applies. Pair `--persistent` with `--no-auto-archive` only for indefinite idle retention,
then archive explicitly.

Codex policy flags belong to the caller. Claude headless defaults to `dontAsk`; bypass only with
explicit authority. Qwen approval remains native and mutable; use `--yolo`, `--no-yolo`, or
`--approval-mode` only as authorized launch preferences. Read [references/contract.md](references/contract.md) for product-specific
events and lifecycle differences.

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 →