mistral-vibe

Delegate multi-step repository tasks to Mistral Vibe CLI with explicit prompts.

5|Updated Apr 23, 2026
One-click install
npx skills add https://github.com/amirkiarafiei/subagent-cli-skills --skill mistral-vibe
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: mistral-vibe
Source: https://github.com/amirkiarafiei/subagent-cli-skills/tree/main/skills/mistral-vibe
Command: npx skills add https://github.com/amirkiarafiei/subagent-cli-skills --skill mistral-vibe

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

It offloads large multi-step repository tasks to Mistral Vibe CLI so your main orchestrator can stay lean while still getting substantial edits, exploration, and implementation done.

Core Features & Use Cases

  • Long-horizon subagent execution: Runs a separate bounded pass for multi-file work, batch edits, refactors, and deep exploration.
  • Context handoff for alignment: Prompts explicitly include goal, prior decisions, scope, constraints, verification, and expected output to reduce “split story” failures in isolated sessions.
  • Programmatic, non-interactive workflow: Uses CLI flags (no interactive mode) to execute deterministically for delegation-style usage in headless toolchains.

Quick Start

Use the mistral-vibe skill to delegate a full feature implementation by running: vibe --prompt "GOAL: [your goal] | DECISIONS: [what you already chose] | SCOPE: [paths] | CONSTRAINTS: [rules] | VERIFICATION: [tests/commands] | OUTPUT: [format]" --model [model_name] --output text 2>&1.

Frequently Asked Questions about mistral-vibe

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I delegate large multi-step repository tasks to a headless CLI for subagent execution?

You can delegate large multi-step repository tasks by invoking a headless CLI programmatically with explicit flags for model selection, shared-state prompt contents, and a verification command to ensure deterministic subagent execution.

What is the best way to run broad code refactors without exhausting my main orchestrator's context window?

Running broad code refactors without exhausting context is best achieved by offloading the heavy code-edit workflows to a separate bounded pass, keeping your main orchestrator lean while handling multi-file work.

How does context handoff work when executing isolated sessions for deep repo exploration?

Context handoff for isolated sessions works by explicitly passing the goal, prior decisions, scope, constraints, verification, and expected output in the prompt, which reduces split-story failures during deep repo exploration.

Can I use programmatic non-interactive CLI invocation for batch edits in automated toolchains?

Yes, you can use programmatic non-interactive CLI invocation to execute batch edits deterministically, applying CLI flags instead of interactive mode to support delegation-style usage within automated toolchains.

What limitations exist when passing shared-state prompts for verification tests during task delegation?

A key limitation is that verification tests require an explicit verification command with controlled expected output formatting, meaning ambiguous prompt contents or undefined verification commands will cause delegation failures.