trace

Generate competing hypotheses and rank evidence to triage ambiguous observations.

16|3|Updated Mar 12, 2026
One-click install
npx skills add https://github.com/sehoon787/my-claude --skill trace-sehoon787
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: trace
Source: https://github.com/sehoon787/my-claude/tree/main/skills/omc/trace
Command: npx skills add https://github.com/sehoon787/my-claude --skill trace-sehoon787

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Ambiguity in observed results and complex root causes require evaluating multiple explanations in parallel; this skill provides a structured lane to generate competing hypotheses, gather evidence, and rank explanations.

Core Features & Use Cases

  • Team-mode tracing: automatically spawns multiple tracer lanes to explore distinct hypotheses in parallel.
  • Evidence-for / Evidence-against: collects, compares, and ranks supporting and refuting artifacts for each hypothesis.
  • Structured synthesis: returns a ranked explanation with a critical unknown and a discriminating probe to accelerate resolution.

Quick Start

Provide a trace run for a given observation by generating three distinct hypotheses, collecting evidence for and against each, and returning a ranked synthesis with the best discriminating probe.

Frequently Asked Questions about trace

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

FAQPage Schema
How do I perform root-cause analysis for ambiguous runtime bugs and performance anomalies?

Root-cause analysis for runtime bugs and performance anomalies requires generating competing hypotheses, collecting evidence for and against each, and returning a ranked synthesis with a discriminating probe to resolve the issue.

What is the best way to triage multiple explanations for a complex architectural question?

Triaging multiple architectural explanations involves spawning parallel tracer lanes to explore distinct hypotheses, comparing supporting and refuting artifacts, and ranking explanations by evidence weight to identify the most likely cause.

How does team-mode tracing work when evaluating competing hypotheses?

Team-mode tracing automatically spawns multiple dedicated lanes to explore distinct hypotheses in parallel, collecting explicit evidence for and against each before synthesizing a final ranked conclusion with a critical unknown.

Can I use this structured tracing workflow for performance anomalies without prior evidence collection?

Yes, the structured tracing workflow triages ambiguous performance anomalies by starting from observation, generating three distinct hypotheses, and then systematically collecting evidence for and against each to determine the most likely cause.

What does a ranked synthesis include when resolving ambiguous observations?

A ranked synthesis for ambiguous observations includes the most likely explanation determined by evidence weight, a critical unknown that remains unresolved, and a discriminating probe to accelerate final resolution.

When should I use parallel hypothesis tracing instead of standard debugging?

Use parallel hypothesis tracing when runtime bugs, performance anomalies, or architectural questions present multiple valid explanations that must be weighed simultaneously with explicit evidence for and against each hypothesis.