audit-correctness

Detect and fix logic bugs violating behavioral invariants in software audits.

10|5|Updated Mar 3, 2026
One-click install
npx skills add https://github.com/JHostalek/dotclaude --skill audit-correctness
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: audit-correctness
Source: https://github.com/JHostalek/dotclaude/tree/main/skills/audit-correctness
Command: npx skills add https://github.com/JHostalek/dotclaude --skill audit-correctness

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill identifies behavior that produces observably wrong results or invalid state, including boundary errors, incorrect conditions, data transformations, workflow transitions, concurrency issues, and integration mismatches.

Core Features & Use Cases

  • Behavioral Modeling: Reconstruct user-visible contracts, domain rules, state machines, invariants, and end-to-end behavior before inspecting implementation details.
  • Systematic Correctness Auditing: Examine control flow, data meaning, numerical logic, time, persistence, interfaces, configuration, concurrency, recovery, and domain-specific behavior.
  • Evidence-Based Findings: Classify confirmed defects, worthwhile improvements, and unresolved questions while documenting violating paths, affected outcomes, safe corrections, and regression verification.
  • Use Case: Apply this Skill to audit a payment workflow for rounding errors, duplicate retries, stale state, migration inconsistencies, and incorrect terminal-state transitions before release.

Quick Start

Use the audit-correctness skill to inspect the specified path for logic bugs and produce an evidence-based correctness report with regression tests and safe fixes.

Frequently Asked Questions about audit-correctness

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

FAQPage Schema
How do I find logic bugs in state machines and concurrency workflows before release?

To find logic bugs in state machines and concurrency workflows, reconstruct behavioral contracts and trace end-to-end paths to identify violations of invariants. This systematic correctness audit verifies control flow and data transformations to detect boundary errors and invalid state transitions.

What is the best way to audit payment workflows for rounding errors and duplicate retries?

The best way to audit payment workflows for rounding errors and duplicate retries is to model domain rules and inspect numerical logic alongside persistence behavior. This process classifies evidence-backed findings and generates regression tests to verify safe corrections.

How do I check for boundary cases and data transformation errors in my code?

To check for boundary cases and data transformation errors, perform a systematic correctness audit that examines data meaning and control flow. By tracing end-to-end paths against reconstructed behavioral contracts, you can isolate incorrect conditions and verify fixes with deterministic tests.

Can I use a correctness audit to fix integration mismatches and deployment mode inconsistencies?

Yes, you can use a correctness audit to fix integration mismatches and deployment mode inconsistencies by inspecting interfaces and configurations. The audit traces violating paths, classifies confirmed defects, and documents safe corrections to ensure behavioral invariants remain intact.

When do I need to reconstruct behavioral contracts for a software audit?

You need to reconstruct behavioral contracts for a software audit before inspecting implementation details to establish expected domain rules and end-to-end behavior. This prerequisite step ensures accurate tracing of control flow and state machines to identify invalid state transitions.

Does regression testing work for verifying fixes to concurrency issues and time-based logic?

Yes, regression testing works for verifying fixes to concurrency issues and time-based logic by providing deterministic proof that corrections resolve the defect. It validates that the adjusted control flow and data transformations adhere to the reconstructed behavioral invariants.