error-resolver_skill

Overview

This skill helps debug errors, crashes, and failing commands by guiding a methodical diagnosis and producing minimal, safe fixes. It emphasizes reproducing the issue, isolating a minimal repro, and validating fixes with clear verification steps. The goal is to resolve the root cause while limiting risk and side effects.

How this skill works

The skill walks you through capturing exact error messages and stack traces, then reducing the problem to the smallest failing test or reproduction. It prompts checks for recent changes, dependency and environment mismatches, and incorrect inputs or assumptions. Once the root cause is identified, the skill proposes the smallest corrective change and specifies verification and rollback steps.

When to use it

• When a test, command, or process consistently fails and you need a reproducible diagnosis.
• When production errors occur and you must minimize risk while restoring functionality.
• When a crash or exception stack trace is unclear and root cause needs isolation.
• When making a targeted fix that must include verification and rollback instructions.
• When preparing a bug report or patch that must include a minimal repro and validation steps.

Best practices

• Always reproduce the error locally or in a safe staging environment before changing code.
• Create the smallest possible failing test or script that demonstrates the issue.
• Read stack traces from top to bottom and correlate frames with recent code changes.
• Check environment details: interpreter/runtime version, dependency versions, env vars, and config.
• Document verification steps and any rollback plan for production fixes.

Example use cases

• A unit test started failing after a dependency update — isolate the change, reproduce, and propose a minimal patch.
• A web endpoint intermittently returns 500 — collect logs, reproduce with a small script, and fix input validation or race condition.
• A CLI command crashes on some systems — verify environment differences, add defensive checks, and provide tests.
• A deployment causes an application error in production — identify the offending release, apply a hotfix with rollback steps, and add a regression test.