# Best Of N Ops

> Operational guidance for Cursor's /best-of-n command. Activates when running parallel model attempts and selecting the strongest result.

- Skill: `mike-diff/best-of-n-ops` (Agent Skill)
- Install (CLI): `npx skillmds@latest add mike-diff/best-of-n-ops`
- Raw SKILL.md: https://api.skillmd.com/api/skills/mike-diff/best-of-n-ops/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: AI & ML
- Author: mike-diff (https://skillmd.com/u/mike-diff)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/mike-diff/best-of-n-ops

---


# /best-of-n — Parallel Model Selection

Guidance for using Cursor's parallel attempt feature to improve output quality.

## When to use

- Ambiguous requirements where multiple approaches are viable
- High-stakes code where correctness matters (security, data integrity)
- Creative tasks where solution diversity improves outcomes

## How it works

1. `/best-of-n` spawns N independent attempts at the same prompt
2. Each attempt runs in its own isolated worktree
3. Results are evaluated and the strongest is selected
4. Other attempts are discarded

## Best practices

- Use N=2–3 for most cases; higher N for critical paths
- Write clear evaluation criteria in the prompt
- Review the selected output before merging — selection is heuristic, not perfect