triage-bugs

Coordinate four adversarial passes to investigate cached bug tickets across clusters.

3|2|Updated Apr 10, 2026
One-click install
npx skills add https://github.com/AdamCaviness/agentic-toolkit --skill triage-bugs-adamcaviness
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: triage-bugs
Source: https://github.com/AdamCaviness/agentic-toolkit/tree/main/skills/triage-bugs
Command: npx skills add https://github.com/AdamCaviness/agentic-toolkit --skill triage-bugs-adamcaviness

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Triaging codebase defects across multiple perspectives is complex; this skill coordinates four parallel sub-agents to uncover, verify, and document defects so maintainers can fix them from the report.

Core Features & Use Cases

  • Coordinated, four-cluster defect investigation (Data & State, Security & Auth, Correctness, Silent Failures)
  • Caching and cross-cluster notes to disk for reproducibility and auditability
  • Adversarial 4-pass analysis (Frame, Trace, Falsify, Prove) that builds rigorous evidence
  • Per-cluster ledgers and a unified post-processor for final consolidation

Quick Start

Run a triage-bugs run to start gathering tickets, assigning clusters, and spawning agents for simultaneous analysis.

Frequently Asked Questions about triage-bugs

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

FAQPage Schema
How do I investigate bugs across multiple codebase perspectives simultaneously?

Bug investigation across multiple perspectives is coordinated by spawning four parallel sub-agents, each analyzing a specific defect cluster—Data & State, Security & Auth, Correctness, and Silent Failures—to uncover and document issues simultaneously.

What is an adversarial 4-pass analysis for bug triage?

Adversarial 4-pass bug triage is an investigation method using Frame, Trace, Falsify, and Prove stages to build rigorous evidence for defects, ensuring findings are verified rather than assumed before maintainers attempt fixes.

How do I coordinate bug triage across clusters in diverse codebases?

Coordinating bug triage across diverse codebases is handled by auto-detecting the ticket system, building a project map, and spawning one sub-agent per cluster to analyze cached open tickets and document findings with rigorous evidence.

Can I triage defects across codebases without manually specifying the ticket system?

Yes, triaging defects across codebases does not require manual ticket system specification because the skill auto-detects the ticket system, reads cached open tickets, and automatically builds a project map for analysis.

How are bug triage findings documented for post-processing?

Bug triage findings are documented by generating per-cluster edit files, a final ledger per cluster, cross-cluster notes cached to disk for reproducibility, and a unified triage summary for the post-processor to consolidate.

What is the best way to triage silent failures and security defects in parallel?

The best way to triage silent failures and security defects in parallel is using coordinated sub-agents that independently investigate specific clusters, applying an adversarial 4-pass analysis to verify and document each defect with rigorous evidence.