antithesis-debug

Debug Antithesis test runs interactively through the multiverse debugger browser interface.

Updated Aug 5, 2026
One-click install
npx skills add https://github.com/harivansh-afk/loom-index-e2e --skill antithesis-debug-harivansh-afk
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: antithesis-debug
Source: https://github.com/harivansh-afk/loom-index-e2e/tree/main/skills/antithesis/debug
Command: npx skills add https://github.com/harivansh-afk/loom-index-e2e --skill antithesis-debug-harivansh-afk

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires agent-browser, and includes references (resource) and assets (resource) components.

What problem does it solve? Investigating failures inside Antithesis test runs requires navigating a complex web-based multiverse debugger, running commands in containers, and extracting evidence from a specific moment in virtual time. This Skill automates that browser-driven workflow so you can inspect container filesystems, runtime state, and logs without manual clicking. ## Core Features & Use Cases - Simplified debugger automation: Run bash commands in containers, read files, list directories, and extract artifacts by driving the simplified debugger UI through injected JavaScript runtime methods. - Advanced notebook mode: Switch to the Monaco-based notebook for programmatic inspection, timeline branching, event set queries, and fault injector state checks. - File extraction and download: Extract files from containers at a chosen moment and download them locally for offline analysis. - Use Case: Given an Antithesis debugging-session URL for a failed test, open the session, click the relevant log line to set the moment and container, run ls -la /opt/antithesis/ inside the container, and extract the workload entrypoint script to diagnose the failure. ## Quick Start Open this Antithesis debugging-session URL and inspect the filesystem of the failing container to find why the workload crashed.

Frequently Asked Questions about antithesis-debug

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

FAQPage Schema
How do I debug an Antithesis test run interactively?▼

Open the debugging-session URL with agent-browser, inject the antithesis-debug runtime, and use the simplified debugger to click log lines, run bash commands in containers, and read output. The runtime exposes window.__antithesisDebug methods for all interactions.

How to extract a file from an Antithesis container?▼

Use the simplified.extractFile method with the target file path, wait for the output with waitForNewOutput, then snapshot the output section to get a download link ref and download it to a local directory with agent-browser.

What is the difference between simplified and advanced Antithesis debugger modes?▼

Simplified mode runs shell commands, reads files, and extracts artifacts through a direct UI. Advanced mode provides a Monaco notebook with JavaScript cells for branching timelines, event set queries, and fault injector inspection. Detect the mode with getMode() and switch with switchMode().

Does the Antithesis debugger require authentication?▼

Yes, debugger URLs typically require login. Use the antithesis-triage skill's setup-auth reference for the interactive login flow, and pair sessions with --session-name antithesis so authentication state is shared across debugging runs.

Why is window.__antithesisDebug undefined during a debugging session?▼

The injected runtime is lost after any page navigation or reload. Rerun the combined injection command piping browser-utils.js and antithesis-debug.js into agent-browser eval --stdin, then call waitForReady before issuing further method calls.

When should I use advanced notebook mode instead of simplified mode?▼

Switch to advanced mode only when you need moment.branch() timeline branching, environment.events queries, fault injector state inspection, or multi-cell JavaScript notebook workflows. Simplified mode covers most command execution and file extraction tasks.