distsys-review

Review distributed-systems code diffs for correctness and operability hazards.

1|Updated Jul 31, 2023
One-click install
npx skills add https://github.com/austintheriot/dotfiles --skill distsys-review
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: distsys-review
Source: https://github.com/austintheriot/dotfiles/tree/main/.claude/skills/distsys-review
Command: npx skills add https://github.com/austintheriot/dotfiles --skill distsys-review

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Prevents subtle distributed-systems correctness and operability failures that often slip past tests by reviewing changed code for failure modes that only appear under partial outage, retries, or scale.

Core Features & Use Cases

  • Expert diff-driven review: Analyzes the current branch versus main by default, or a specific path or PR/range you provide, using severity-labeled findings with file:line references.
  • Failure-mode checklist coverage: Focuses on high-frequency production issues like retries without idempotency, unbounded queues, missing fencing tokens, replication-lag races, cache stampedes, and other cross-process hazards.
  • Targeted subagent routing: Delegates deeper runtime vs. data-model questions to the appropriate subagents for more informed guidance, while keeping reporting actionable and non-chatty.

Quick Start

Run distsys-review on your current branch diff so it checks for distributed-systems footguns in the changes.

Frequently Asked Questions about distsys-review

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

FAQPage Schema
How do I review branch diffs for distributed systems correctness hazards?

Analyze branch diffs against main to detect distributed systems correctness hazards. The review checks changed code for failure modes like retries without idempotency, missing fencing tokens, and replication-lag races, providing severity-labeled findings with file:line references.

What common distributed systems failure modes should I look for during a code review?

Distributed systems code reviews should target retries without idempotency, unbounded queues, missing fencing tokens, replication-lag races, and cache stampedes. These cross-process hazards often slip past standard tests and fail under partial outage or scale.

Can I target a specific file path or PR number for a distributed systems code review?

Yes, distributed systems code review supports targeted file paths, PR numbers, and git ranges. By default it analyzes the current branch diff against main, allowing you to scope the review to specific changes.

How does this code review handle partial failure scenarios in distributed systems?

It reviews code changes to detect operability hazards likely to fail under partial failure, retries, or scale. The process applies category-based findings with severity labels and delegates deeper runtime or data-model questions to specialized subagents.

Does the distributed systems review automatically post comments to my pull requests?

No, the distributed systems review avoids comment posting. It generates actionable, non-chatty reports with file:line references and severity labels, keeping the feedback contained to the review workflow.

When do I need to check for idempotency and fencing tokens in distributed systems code?

You need to check for idempotency and fencing tokens whenever changed code involves retries or cross-process coordination. Reviewing these mechanisms early prevents subtle correctness failures that only appear during partial outages or network retries.