home / skills / hiroro-work / claude-plugins / ask-peer

ask-peer skill

safe

This skill helps you validate plans, review code, and brainstorm with a peer to ensure correctness and uncover edge cases.

npx playbooks add skill hiroro-work/claude-plugins --skill ask-peer

Review the files below or copy the command above to add this skill to your agents.

Files (1)

SKILL.md

1.6 KB

---
name: ask-peer
description: Consult with a peer engineer for plan review, code review, implementation discussions, or problem-solving brainstorming. Use when you need a second opinion, want to validate your approach, or check for overlooked issues.
---

# Peer Engineer Consultation

Get a second opinion from a peer engineer (Claude subagent) for:

- **Planning review**: Validate your implementation approach before starting
- **Code review**: Check completed work for issues or improvements
- **Problem-solving**: Brainstorm solutions when stuck
- **Sanity check**: Confirm you're on the right track

## Usage

Use the `/ask-peer` command followed by your consultation request:

```bash
/ask-peer Review this implementation plan for adding user authentication
```

Or simply describe what you need help with:

```bash
/ask-peer I'm stuck on how to handle error cases in the payment flow
```

## Examples

**Plan review before implementation:**

```bash
/ask-peer Review my plan to refactor the authentication module. Check for security concerns and missing edge cases.
```

**Code review after completion:**

```bash
/ask-peer Review the changes I made to src/auth/. Look for bugs, security issues, and code quality problems.
```

**Problem-solving consultation:**

```bash
/ask-peer I'm getting intermittent test failures in the payment module. Help me think through possible causes.
```

## What the Peer Agent Provides

- Frank, objective feedback as an equal
- Specific concerns with suggested alternatives
- Questions to challenge assumptions
- Practical solutions rather than perfection

Overview

This skill connects you with a peer engineer subagent for plan reviews, code reviews, implementation discussions, and brainstorming. It provides candid, constructive feedback to validate approaches, spot gaps, and suggest practical fixes. Use it when you want a second opinion or to catch overlooked issues before they become problems.

How this skill works

Invoke the skill with a concise prompt describing what you need reviewed or discussed. The peer agent reads your plan, code snippet, or problem description and returns frank, objective feedback: specific concerns, alternative approaches, and targeted follow-up questions. Responses focus on actionable advice—security, edge cases, testing, and trade-offs—rather than vague commentary.

When to use it

Before starting an implementation to validate the overall plan and highlight risks
After completing work for a focused code review (bugs, security, style, performance)
When stuck on a bug or intermittent failure to brainstorm potential root causes
To sanity-check assumptions, system boundaries, or design trade-offs
To prepare for a design doc review or engineering interview feedback

Best practices

Give a short context summary, goals, and constraints up front
Include key code snippets, config, logs, or the plan outline—avoid sending entire repos
State the level of feedback desired (quick sanity check, deep review, alternatives)
Mention nonfunctional priorities (security, latency, test coverage) to focus replies
Be explicit about timelines or blockers so suggestions are practical

Example use cases

Review a planned refactor of the authentication module for missing edge cases and security gaps
Inspect recent changes in src/auth/ for bugs, insecure patterns, or performance issues
Troubleshoot intermittent test failures in the payment flow by brainstorming likely causes
Validate error-handling strategy across async workflows and suggest improvements
Get alternate implementation ideas when choosing between complexity and maintainability

FAQ

How detailed should my prompt be?

Provide enough context to make feedback actionable: goals, constraints, key code or logs, and what you want reviewed.

Will the agent rewrite my code?

The agent suggests concrete changes and examples but focuses on review and alternatives rather than full rewrites unless requested.