correctness-reviewer

Trace code execution to detect logic, boundary, race condition, and error-propagation bugs, outputting JSON findings per schema.

Updated Aug 23, 2026
One-click install
npx skills add https://github.com/xbpk3t/ce-codex --skill correctness-reviewer
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: correctness-reviewer
Source: https://github.com/xbpk3t/ce-codex/tree/main/skills/correctness-reviewer
Command: npx skills add https://github.com/xbpk3t/ce-codex --skill correctness-reviewer

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Reviews code to find logic errors, boundary mistakes, state-transition faults, race conditions, and broken error propagation that unit tests and style checks often miss, helping prevent silent failures and incorrect behavior in production.

Core Features & Use Cases

  • Execution tracing: Mentally execute code paths to expose off-by-one errors, incorrect branching, and boundary math mistakes.
  • State and concurrency analysis: Identify incorrect state transitions, flags left in erroneous states, and potential race conditions or TOCTOU issues.
  • Error propagation checks: Detect swallowed errors, mis-mapped error codes, and fallback behavior that masks failures.
  • Use Case: Apply this reviewer to pull requests or critical modules to surface reproducible correctness issues before release.

Quick Start

Run a correctness review on the attached code and return findings as JSON matching the specified schema.

Frequently Asked Questions about correctness-reviewer

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

FAQPage Schema
How do I find race conditions and state management bugs in a pull request?

To detect race conditions and state management bugs in a pull request, review the code by tracing control flow and checking state transitions. This surfaces concurrency issues and flags left in erroneous states before release.

What is the best way to check error propagation and find swallowed exceptions in code?

The best way to check error propagation is to trace failure paths and identify swallowed exceptions or mis-mapped error codes. This detects fallback behavior that masks failures and prevents silent issues in production.

How do I trace execution paths to find off-by-one errors and boundary math mistakes?

Trace execution paths by mentally executing code branches to expose off-by-one errors and boundary math mistakes. This execution tracing identifies incorrect branching and validates behavioral correctness in critical modules.

Can I review code snippets for logic errors and output findings as JSON?

Yes, you can review code snippets for logic errors and output findings as JSON. The review traces control flow and checks boundary conditions, returning reproducible correctness issues matching a specified JSON schema.

Why does my code pass unit tests but still have silent failures in production?

Code passes unit tests but fails silently in production due to subtle logic errors, broken error propagation, and race conditions that standard tests miss. Reviewing state transitions and error mapping exposes these hidden behavioral faults.