# Ab Compare

> Generate and compare multiple implementation variants with benchmarking and scoring

- Skill: `chainlesschain/ab-compare` (Agent Skill, multi-file: 2 files)
- Install (CLI): `npx skillmds@latest add chainlesschain/ab-compare`
- Raw SKILL.md: https://api.skillmd.com/api/skills/chainlesschain/ab-compare/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Web & Frontend
- Author: chainlesschain (https://skillmd.com/u/chainlesschain)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/chainlesschain/ab-compare

---


# A/B Compare

## Description

Generates multiple implementation variants for a task using different agent profiles, benchmarks each variant, and selects the best one.

## Usage

```
/ab-compare "task description" [--variants 3] [--benchmark]
```

## Agent Profiles

| Profile          | Strategy                                   |
| ---------------- | ------------------------------------------ |
| concise-agent    | Minimal code, one-liners, built-in methods |
| robust-agent     | Comprehensive error handling, edge cases   |
| readable-agent   | Self-documenting, clarity over brevity     |
| performant-agent | Optimized, minimal allocations             |
| testable-agent   | Pure functions, DI, clear interfaces       |

## Scoring

| Criterion      | Weight | Description                   |
| -------------- | ------ | ----------------------------- |
| Conciseness    | 30%    | Fewer lines = higher score    |
| Error Handling | 35%    | try/catch, null checks, throw |
| Readability    | 35%    | Line length, comments, naming |

## Output Format

```
A/B Comparison Report
=====================
Task: "add input validation to login form"
Variants: 3

| Agent          | Concise | Error | Readable | Total |
| -------------- | ------- | ----- | -------- | ----- |
| concise-agent  | 90      | 50    | 60       | 65    |
| robust-agent   | 40      | 95    | 70       | 78    |
| readable-agent | 60      | 65    | 90       | 73    |

Winner: robust-agent (78/100)
```