pony-code-review

Name: pony-code-review
Availability: InStock
Author: Pony

Orchestrate multi-persona code reviews on PRs, branches, or local changes.

6|Updated Mar 9, 2026

Search Tags:#workflow #personas #ai-assistant #code-review #pr-analysis #ensemble

Ready to run

npx skills add https://github.com/ponylang/llm-skills --skill pony-code-review

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: pony-code-review
Source: https://github.com/ponylang/llm-skills/tree/main/pony-code-review
Command: npx skills add https://github.com/ponylang/llm-skills --skill pony-code-review

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Ensemble code review that synthesizes findings from specialized reviewer personas. Has two modes: full (8 personas, iterative re-review) and lightweight (3 personas, single pass). Not for one-line config changes or typo fixes — ask permission to skip review for those.

Core Features & Use Cases

Orchestrated multi-persona analysis for PR reviews and branch changes.
Supports full and lightweight modes with escalation loops and synthesis.
Integrates with /pony-ensemble and /pony-synthesize to produce a final consolidated report.

Quick Start

Invoke the skill on a PR to run a full-ensemble review and generate the findings report.

Frequently Asked Questions about pony-code-review

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

FAQPage Schema

How do I use AI personas for ensemble code review on a PR?▼

Ensemble code review uses multiple specialized AI reviewer personas to analyze a PR, branch, or local changes. It orchestrates analyses for correctness, adversarial issues, and tests, then synthesizes evidence artifacts into a consolidated feedback report.

What's the best way to review complex code changes across multiple subsystems?▼

For complex code changes across subsystems, a full ensemble review applies 8 personas with iterative re-review and escalation loops. This generates structured feedback and evidence synthesis to catch issues standard reviews might miss.

When should I use lightweight mode instead of a full multi-persona code review?▼

Lightweight mode uses 3 personas in a single pass for faster analysis. Use it for smaller changes, while full mode handles complex subsystems. For one-line config changes or typo fixes, skip review entirely.

Does AI-assisted code review require specific workflow orchestrators?▼

AI-assisted ensemble code review requires the /pony-ensemble orchestrator and /pony-synthesize workflow to guide mode selection, manage escalation, and produce the final consolidated findings report.

How do AI reviewer personas synthesize findings into a single PR analysis?▼

Reviewer personas generate individual analyses with evidence artifacts. The synthesis workflow then consolidates these findings, escalating issues when necessary, to produce a final structured PR review report.

Can I automate PR analysis for branch changes without manual intervention?▼

You can invoke the ensemble code review skill on a PR or branch changes to automatically run multi-persona analysis. It guides mode selection and escalation to produce a consolidated findings report autonomously.