debug-mode

github:nicobailon/debug-mode · first committed 2025-12-12 · source · badge

Hypothesis-driven debugging with hybrid dual-track parallel execution (Opus 4.5 + GPT 5.2). Spawns two independent chains of subagents where each reviews and improves upon its own previous work, then synthesizes findings from both tracks. Use when debugging hard-to-reproduce bugs, CI/E2E test failures, flaky tests, or when standard fixes have failed.

Skill Health: not yet scored

Listed, not yet scored. This skill has no public model test yet. It is in the scoring queue; submitted skills go first, then the most-installed.

Version history

10 commits to this file on GitHub (newest first).

2025-12-12 · commit 252cb4c · Add GPT 5.2 behavior constraints to prompts
2025-12-12 · commit bcffc58 · Fix cleanup false positives + add repomix quick reference
2025-12-12 · commit b4c1547 · Add Judge subagent to compare tracks and pick winner
2025-12-12 · commit f37c80f · Allow A2/A3 and B2/B3 to verify and exit early
2025-12-12 · commit 2e41e5f · Track A: chain A1-A3 with resume, GPT for final verify
2025-12-12 · commit 2ea0f45 · Track 0 Orchestrator + Opus 4.5 everywhere
2025-12-12 · commit a57d494 · Require repomix for context bundling in Track 0
2025-12-12 · commit 04dfcb6 · Add Context Builder as first step of Track 0
2025-12-12 · commit cd5141f · Major architecture overhaul: Track 0, alternating models, fix-attempt workflow
2025-12-12 · commit 84027c2 · Initial release: debug-mode skill for Claude Code

Archived snapshots