Steel Session Debugger

•
2
•

Diagnoses failed web automation sessions using browser logs, semantic traces, and network forensics to provide evidence-backed fixes.

This skill provides a specialized diagnostic framework for troubleshooting failed Steel browser automation sessions. It moves beyond guesswork by systematically collecting and analyzing session metadata, browser console logs, raw agent traces, and network activity. By reconstructing a timeline of events and redacting sensitive PII, it enables developers to identify the root cause of timeouts, blocked pages, and unexpected agent behaviors, delivering a structured diagnosis and clear verification steps for every failure.

Key Features

01Timeline reconstruction to visualize session creation, navigation, and specific failure points
022 GitHub stars
03Automated PII redaction for sensitive data like cookies, tokens, and form values
04Integrated CLI diagnostics via Steel CLI for deep inspection of automation runs
05Comprehensive evidence collection including browser logs, semantic traces, and session metadata
06Failure classification based on a standardized taxonomy for consistent reporting

Use Cases

01Troubleshooting intermittent timeouts or browser errors in automated scraping scripts
02Analyzing network-level failures to determine if specialized proxy or anti-bot measures are required
03Diagnosing why an agent failed to navigate or interact with specific page elements