investigate

Diagnose bugs through systematic root cause investigation before applying fixes.

Updated Apr 5, 2026
One-click install
npx skills add https://github.com/Simon-YHKim/eject-button --skill investigate-simon-yhkim
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: investigate
Source: https://github.com/Simon-YHKim/eject-button/tree/main/.claude/skills/investigate
Command: npx skills add https://github.com/Simon-YHKim/eject-button --skill investigate-simon-yhkim

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Debugging often devolves into guessing and symptom-patching, which creates whack-a-mole bugs and regressions. This Skill enforces a disciplined four-phase debugging workflow that finds the actual root cause before any code is changed. ## Core Features & Use Cases - Root Cause First (Iron Law): No fixes are applied until a specific, testable root cause hypothesis is formed from symptoms, stack traces, code tracing, and git history. - Four-Phase Workflow: Investigate, analyze patterns (race conditions, null propagation, state corruption, config drift), test hypotheses with a 3-strike escalation rule, then implement a minimal fix with a regression test. - Scope Locking: Optionally freezes edits to the affected directory so debugging does not creep into unrelated code. - Use Case: A user reports "the checkout endpoint returns 500 intermittently." The Skill traces the code path, checks recent commits, identifies a race condition on shared state, confirms it with a temporary log, then ships a minimal fix plus a failing-then-passing regression test. ## Quick Start Ask the assistant to investigate why the login flow throws a 500 error after the latest deploy.

Frequently Asked Questions about investigate

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

FAQPage Schema
How do I debug a bug systematically instead of guessing?

Follow a four-phase process: collect symptoms and reproduce the issue, trace the code path and check recent git changes, form and test one hypothesis at a time, then implement a minimal fix with a regression test. Never patch before confirming the root cause.

What is root cause analysis for software bugs?

Root cause analysis identifies the underlying defect causing a symptom rather than patching surface behavior. It involves reading stack traces, tracing data flow, reviewing recent commits, and confirming a testable hypothesis before writing any fix.

How do I debug intermittent or timing-dependent errors?

Intermittent failures often indicate race conditions on shared state or stale caches. Add temporary logging at suspected points, reproduce deterministically if possible, and check for concurrent access patterns before proposing a fix.

When should debugging escalate to a human instead of continuing?

Escalate after three failed hypotheses, when a fix touches more than five files, or when the change is security-sensitive and cannot be verified. Repeated failures usually signal an architectural problem rather than a simple bug.

Why does a bug keep recurring in the same files?

Recurring bugs in the same area are an architectural smell, not a coincidence. Check git history for prior fixes in those files and consider whether the module's design, rather than individual lines, is the root cause.