bug-catcher

Diagnose software bugs with adversarial verification and severity-ranked fix plans.

1|Updated Jun 17, 2026
One-click install
npx skills add https://github.com/Sfzmango/Maungs-agentic-toolbelt --skill bug-catcher
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: bug-catcher
Source: https://github.com/Sfzmango/Maungs-agentic-toolbelt/tree/main/plugins/maungs-agentic-toolbelt/skills/bug-catcher
Command: npx skills add https://github.com/Sfzmango/Maungs-agentic-toolbelt --skill bug-catcher

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Manual bug triage often misses root causes or misprioritizes issues, leading to wasted effort on incorrect fixes or critical bugs slipping through production. This skill eliminates that risk by using a dual-agent adversarial process to confirm diagnoses before handing off vetted, severity-ranked fix plans.

Core Features & Use Cases

  • Targeted single-bug diagnosis: For a specific reported bug, delegates root cause analysis to a dedicated subagent, runs an adversarial double-check to confirm the diagnosis, and produces an evidence-backed fix plan.
  • Full codebase-wide sweeps: With the --global flag, scans all code, tests, E2E playbooks, and documentation for bugs, verifies each candidate adversarially, and outputs a severity-ranked backlog.
  • Cosmetic doc auto-remediation: Automatically fixes unambiguous, purely presentational documentation defects (like broken mermaid blocks or malformed tables) during global sweeps without manual intervention.
  • Use Case: If your team is seeing intermittent 500 errors on the checkout flow, use this skill to diagnose the root cause, verify it is not a false positive, and get a prioritized fix plan routed to the implementation team.

Quick Start

Invoke the bug-catcher skill with a clear description of the bug you are experiencing, or add the --global flag to run a full sweep of your entire codebase for all potential bugs.

Frequently Asked Questions about bug-catcher

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

FAQPage Schema
How does adversarial verification improve bug diagnosis accuracy?

Adversarial verification improves bug diagnosis by running a dual-agent double-check process to confirm root cause analysis before producing an evidence-backed fix plan. This mandatory cross-examination prevents misdiagnosed issues and eliminates false positives during bug triage.

How do I run a full codebase-wide sweep for all potential bugs?

To run a full codebase-wide sweep for all potential bugs, invoke the diagnostic skill with the --global flag. This scans all application code, test suites, E2E playbooks, and project documentation, outputting a severity-ranked bug backlog after adversarially verifying each candidate.

Can I automatically fix broken documentation defects during a global codebase audit?

Yes, you can automatically fix broken documentation defects during a global codebase audit. The sweep auto-remediates purely presentational documentation issues like malformed tables and broken mermaid blocks without requiring manual intervention.

What is the best way to prioritize intermittent software bugs for the implementation team?

The best way to prioritize intermittent software bugs is using severity-ranked bug backlogs generated from confirmed root cause analysis. This gated handoff provides the implementation pipeline with evidence-backed, prioritized fix plans for controlled execution.

Does this bug triage process work for single reported bugs or only full codebase sweeps?

This bug triage process works for both single reported bugs and full codebase sweeps. For single-bug triage, it delegates targeted root cause analysis to a dedicated subagent and runs an adversarial double-check before routing the fix plan.