skills.rest
marcocello
marcocelloCOMMUNITY
Explore all skills by this author →
WebsiteGitHub

coding-repair

Diagnose the latest failure and apply the smallest verified fix.

10|1|Updated Feb 22, 2026
Search Tags:#testing#debugging#verification#failure analysis#regression#repair
One-click install
npx skills add https://github.com/marcocello/dot-codex --skill coding-repair
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: coding-repair
Source: https://github.com/marcocello/dot-codex/tree/main/skills/coding-repair
Command: npx skills add https://github.com/marcocello/dot-codex --skill coding-repair

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill helps repair a known defect or concrete failure by quickly identifying the owning problem, making the smallest correct change, and verifying the fix at the narrowest meaningful boundary.

Core Features & Use Cases

  • Fast failure triage: Reads the latest error signal, classifies ownership, and focuses on the relevant code, setup, proof, or tooling path.
  • Minimal local repair: Updates the smallest necessary surface area while preserving accepted behavior and avoiding unnecessary refactors.
  • Narrow verification: Re-runs the targeted check first, then broader proof when needed, so the fix is confirmed before handoff.
  • Use Case: A regression appears in a build, test, browser flow, or evaluator run, and the Skill is used to isolate the cause and apply a verified fix without widening scope.

Quick Start

Use the coding-repair skill to inspect the latest failure, identify the owning component, and apply the smallest verified fix.

Frequently Asked Questions about coding-repair

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

FAQPage Schema
How do I fix a regression failure with minimal code changes?▼

Fix a regression failure by reading the latest error signal, classifying ownership, and updating the smallest necessary code surface. The targeted check is re-run first, then broader proof, confirming the fix before handoff.

What is the best way to triage a failing command or runtime issue?▼

Triage a failing command by reading the latest error, classifying ownership across code or tooling, and applying a minimal local repair. This preserves accepted behavior and avoids unnecessary refactors during failure analysis.

Does this debugging approach work for evaluator findings and setup errors?▼

Yes, this debugging approach applies to evaluator findings, setup errors, and failing commands. It diagnoses the owning problem across code, setup, proof, and tooling paths, then applies the smallest correct change to resolve the defect.

How do I verify a bug fix before running the full test suite?▼

Verify a bug fix by re-running the targeted check at the narrowest meaningful boundary first. Once the specific failure is confirmed repaired, execute broader proof or full test runs only when needed before handoff.

Can I use this method to repair tooling and proof failures without widening scope?▼

Yes, you can repair tooling and proof failures without widening scope. The method updates the smallest necessary surface area, avoiding unnecessary refactors while ensuring the fix is verified at the narrowest boundary before broader reruns.

Related Skills

Explore More Software Engineering Skills→
adamatdevopsadamatdevops
1

error-analysis

Analyze LLM pipeline traces to prioritize failure modes by root cause.

Community
Intermediate
robkamrobkam
12

bugfix-red-green-proof

Enforce a red-green-proof loop starting with a failing regression test.

Community
Advanced
sunghwanismsunghwanism

verification-engine

Automate build, test, and lint workflows with failure analysis and fix suggestions.

Community
Intermediate
robkamrobkam
12

qa-root-cause-remediation

Diagnose QA gate failures and implement root-cause fixes with rerun plans.

Community
Advanced