Verified:GitHub

debug-e2e

Debug failing e2e tests by delegating log analysis to subagents.

Updated May 14, 2026

Search Tags:#log analysis #flaky tests #hypothesis testing #e2e testing #test orchestration #ci debugging #multi-node systems

Ready to run

npx skills add https://github.com/HabibTorjmen/Blockchain --skill debug-e2e-habibtorjmen

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: debug-e2e
Source: https://github.com/HabibTorjmen/Blockchain/tree/main/aztec-packages/yarn-project/.claude/skills/debug-e2e
Command: npx skills add https://github.com/HabibTorjmen/Blockchain --skill debug-e2e-habibtorjmen

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Failed or flaky end-to-end (e2e) tests are hard to debug because logs are extremely large and interpreting them inline pollutes the conversation context.

Core Features & Use Cases

Interactive e2e debugging orchestration: Guides a ping-pong debugging session by coordinating steps without directly consuming raw logs.
Delegated log analysis via subagents: Uses dedicated subagents to identify CI failures and deep-dive specific log regions, keeping the main conversation focused.
Hypothesis-driven workflow: Helps form competing theories, verify evidence, compare failing vs passing runs, and then pivot to codebase investigation.

Quick Start

Run /debug-e2e 343c52b17688d2cd to identify the failing e2e tests for that CI hash and walk through hypotheses with log analysis delegated to subagents.

Frequently Asked Questions about debug-e2e

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

FAQPage Schema

How do I debug failing e2e tests without overwhelming the conversation context?▼

To debug failing e2e tests without context bloat, this Skill orchestrates a hypothesis-driven workflow that delegates raw log reading and deep analysis to subagents, keeping the main conversation focused on failure identification and evidence verification.

How do I investigate flaky tests by comparing passing and failing CI runs?▼

Investigate flaky tests by forming competing hypotheses and using delegated subagents to deep-dive specific log regions. The workflow supports comparing failing versus passing runs to verify evidence and pinpoint timeout or assertion discrepancies.

Can I trigger e2e test debugging using a CI hash or PR number?▼

Yes, you can trigger e2e test debugging by providing a CI hash, PR number, CI URL, or specific test name. The workflow identifies the failing tests from these inputs and initiates a guided, interactive debugging session.

What is the best way to orchestrate CI debugging for multi-node systems?▼

The best way to orchestrate CI debugging for multi-node systems is using a ping-pong debugging session that coordinates steps without directly consuming raw logs. This approach manages hypothesis refinement and drives optional codebase investigation.

Does e2e test debugging support local test reproduction after log analysis?▼

Yes, e2e test debugging supports local test reproduction. After identifying the specific failure, analyzing relevant log evidence via subagents, and refining hypotheses, the workflow drives optional codebase investigation and local reproduction steps.