debugging

Diagnose defects systematically using evidence-driven, technology-agnostic debugging practices.

SKILL.md

GitHub

debugging

Use this skill for failures, exceptions, incorrect behavior, intermittent bugs, and root-cause analysis.

Principles

Reproduce before changing code when practical.

Separate:

  • symptom
  • trigger
  • root cause

Collect evidence from relevant:

  • errors
  • logs
  • inputs
  • state
  • recent changes
  • traces
  • database behavior
  • network behavior

Reduce the problem to the smallest useful reproduction.

Avoid broad speculative changes.

Once root cause is identified:

  1. make the smallest correct fix
  2. add regression protection where useful
  3. verify nearby behavior
  4. remove temporary debugging instrumentation

Do not hide failures merely by suppressing exceptions or logs.

Adaptation

Use project evidence to determine the actual language, framework, runtime, and existing conventions.

When stack-specific implementation guidance is needed, prefer project evidence, native framework or language mechanisms, and relevant user-installed specialist skills. Technology-specific guidance must not redefine or weaken the core engineering requirement.