Minimax Runner

Name: Minimax Runner
Author: robsonrung

robsonrung/rar-skills

30 installs
1 repo stars
Updated August 4, 2026
robsonrung/rar-skills

Executes prompts against Minimax models through the Qwen Code CLI in headless mode, adding a Minimax seat for councils and scripted workflows.

About

A runner skill that routes prompts to Minimax models via the shared Qwen CLI wrapper. A developer uses it to add a Minimax-backed perspective to cross-runner councils without leaving the workspace.

Runs Minimax models through the shared Qwen CLI wrapper, headless and stream-friendly
Requires a Minimax provider configured in the qwen CLI settings

Minimax Runner by the numbers

30 all-time installs (skills.sh)
+6 installs in the week ending Aug 5, 2026 (Skillselion tracking)
Ranked #9,307 of 16,546 AI & Agent Building skills by installs in the Skillselion catalog
Data as of Aug 5, 2026 (Skillselion catalog sync)

npx skills add https://github.com/robsonrung/rar-skills --skill minimax-runner

Add your badge

Show developers this skill is listed on Skillselion. Paste this into your README.

[![Listed on Skillselion](https://skillselion.com/badge/skills/robsonrung/rar-skills/minimax-runner.svg)](https://skillselion.com/skills/robsonrung/rar-skills/minimax-runner)

Installs	30
repo stars	★ 1
Last updated	August 4, 2026
Repository	robsonrung/rar-skills ↗

What it does

Executes prompts against Minimax models through the Qwen Code CLI in headless mode, adding a Minimax seat for councils and scripted workflows.

Files

SKILL.mdMarkdownGitHub ↗

Minimax Runner

Execute prompts against Minimax models through the shared Qwen CLI wrapper. This skill adds a Minimax-backed seat for councils and scripted workflows while keeping execution headless and stream-friendly.

Default Model

minimax/minimax-m2.7

Pass --model if you want to target another Minimax model exposed by the local qwen CLI account.

Prerequisites

qwen installed and in PATH
A Minimax provider configured in the qwen CLI — a modelProviders entry in ~/.qwen/settings.json (with its API key env var set) or credentials supplied via --openai-api-key / --auth-type. The legacy qwen auth subcommand has been removed.

Security Model

This skill delegates to qwen-runner, so it has the same execution and data sharing model as the Qwen wrapper. Prompt text, prompt files, session files, metadata, and any files Qwen reads during the run may be sent to the selected Minimax provider. Analysis roles (every role except implementer) default to restricted mode (read-only overlay plus plan approval mode); pass --allow-write to opt out. Otherwise approval mode defaults to default.

Shared Wrapper Reference

Supported options, roles, the --json output envelope key contract, return codes, and gotchas are identical to the shared wrapper — read the qwen-runner skill's SKILL.md (../qwen-runner/SKILL.md) when you need flag or envelope details. The envelope is produced by qwen-runner/scripts/run_qwen.py.

Usage

python3 .agents/skills/minimax-runner/scripts/run_minimax.py "your prompt here"

Examples

python3 .agents/skills/minimax-runner/scripts/run_minimax.py "Stress-test this implementation idea"
python3 .agents/skills/minimax-runner/scripts/run_minimax.py --prompt-file /tmp/review.md --role challenger
python3 .agents/skills/minimax-runner/scripts/run_minimax.py "Return JSON only" --output-format json --json

Behavior

1. Delegates to the shared qwen-runner implementation with runner identity set to minimax. 2. Uses stream-json as the default native Qwen output format. 3. Never falls back to another provider. If the run fails (non-zero return_code or auth_ok=false in the envelope), treat the Minimax seat as blocked and report it unavailable so councils can account for the missing seat — never substitute another provider. 4. Preserves the shared wrapper envelope so councils can compare Minimax output with other runners consistently.

# Consumed by the Codex app: exposes this skill as a native subagent seat with UI metadata and a default prompt (see the claude-runner and codex-mission-control skills' SKILL.md notes). Do not remove.
interface:
  display_name: "Minimax Runner"
  short_description: "Run workspace prompts through Qwen CLI on Minimax"
  default_prompt: "Use $minimax-runner to execute this task with Minimax from the current workspace."

#!/usr/bin/env python3
"""Execute prompts against Minimax models through Qwen CLI headless mode."""

import sys
from pathlib import Path


QWEN_RUNNER_DIR = Path(__file__).resolve().parents[2] / "qwen-runner" / "scripts"
sys.path.insert(0, str(QWEN_RUNNER_DIR))

import run_qwen  # noqa: E402


DEFAULT_MODEL = "minimax/minimax-m2.7"


if __name__ == "__main__":
    run_qwen.main(
        default_model=DEFAULT_MODEL,
        runner_name="minimax",
        description="Execute prompts using Qwen CLI against Minimax models in headless mode.",
    )