Crash Loop Investigation
A systematic debugging loop designed to identify, reproduce, and eliminate application crashes. This loop guides you through iterative analysis and fix attempts until the crash is fully resolved.
Goal
Eliminate crashes
How to Run
Input crash details and let the agent iteratively investigate and resolve the issue.
- 01
Provide Crash Context
Share the error message, stack trace, and application name where the crash occurs.
- 02
Reproduction Steps
Describe how to trigger the crash consistently.
- 03
Code Reference
Include relevant code snippets or file paths related to the crash.
Workflow Steps
- 01
Analyze Crash Data
Examine the provided error message and stack trace to identify potential root causes.
- 02
Reproduce the Crash
Run the application following the given reproduction steps to confirm the crash behavior.
- 03
Identify Root Cause
Use debugging techniques to pinpoint the exact code location and reason for the crash.
- 04
Propose Fix
Suggest a targeted code modification to resolve the identified issue.
- 05
Apply Fix
Implement the proposed change in the appropriate file(s).
- 06
Test Fix
Run the crash reproduction again to verify if the fix resolves the issue.
- 07
Evaluate Exit Condition
If the crash no longer occurs, exit the loop; otherwise, iterate with new insights.
Kickoff Prompt
Start the "Crash Loop Investigation" loop. Goal: Eliminate crashes Max iterations: 10 Between iterations run: crash reproduction Exit when: Crash no longer reproducible To begin crash investigation, please provide the following: 1. The exact error message or exception details 2. Application name and version 3. Step-by-step instructions to reproduce the crash 4. Any relevant code snippets or file paths related to the crash 5. Environment details (OS, browser, device, etc.) Once shared, I'll analyze the crash and attempt to resolve it through iterative testing. Self-pace this loop. After each iteration, run `crash reproduction` and evaluate the output, and only continue if the exit condition is not met (Crash no longer reproducible). Stop when the exit condition passes or 10 iterations are reached. Give a short status update each pass.
Guardrails
hardcoded- ·Only modify code directly related to the identified crash.
- ·Do not exceed 10 iterations without progress.
- ·Ensure each code change is minimal and testable.
- ·Document every change made during the loop.
Flow Diagram
Related loops — Debugging
Debugging
Reproduce and Fix
This loop guides you through reproducing a reported bug, identifying its root cause, implementing a fix, and verifying the solution through automated testing. The agent will iteratively work to resolve the issue while maintaining system integrity.
Debugging
Error Log Reduction
This loop analyzes application error logs to identify and fix recurring errors, reducing their frequency over time through iterative debugging and targeted code improvements.
Debugging
Production Incident Resolver
A coding agent loop designed to diagnose and resolve production incidents through iterative investigation, targeted fixes, and continuous health monitoring until system stability is restored.