deep-debug

✓Verified·Scanned 2/18/2026

Multi-agent investigation for stubborn bugs. Use when: going in circles debugging, need to investigate browser/API interactions, complex bugs resisting normal debugging, or when symptoms don't match expectations. Launches parallel agents with different perspectives and uses Chrome tools for evidence gathering.

by jezweb·v10a1f16·29.6 KB·162 installs
Scanned from main at 10a1f16 · Transparency log ↗
$ vett add jezweb/claude-skills/deep-debug

Deep Debug Skill

Multi-agent investigation pattern for stubborn bugs that resist normal debugging.

Origin

This skill emerged from a real debugging session where:

  • We went in circles for ~30 minutes guessing at causes (caching? stream parsing?)
  • User showed Network tab screenshot revealing 2 fetch requests (the actual evidence)
  • Launched 3 parallel agents that found the root cause in ~5 minutes
  • Fix was a React useCallback dependency issue

Key Insight

Evidence before hypotheses. The Network tab screenshot broke the loop. We should have gathered that evidence earlier using Chrome MCP tools.

When to Invoke

  • "I'm stuck on this bug"
  • "We've tried multiple things and nothing works"
  • "This is going in circles"
  • "Can you do a deep investigation?"
  • /deep-debug

Requirements

  • Chrome MCP tools (for browser bugs) - mcp__claude-in-chrome__*
  • Access to spawn sub-agents: debugger, code-reviewer, Explore

See Also

  • Built-in debugger agent - Single-agent debugging
  • skills/agent-development/SKILL.md - Agent patterns and when to delegate