j-debug

Prompt file imported from jlaws/dotfiles (.codex/prompts/j-debug.md). Fill in {{arguments}} before use. Copyright stays with the author.

Bug / failing test: {{arguments}}

Before investigating, gather diagnostic context:

1. Check git status for uncommitted changes (note them for context).
2. Detect test runner from project config (package.json scripts, pytest.ini, Makefile, etc.).
3. Check recent commits (git log --oneline -10) for potential culprits.
4. Capture failure output: If {{arguments}} references a test name, run it first to get the current failure output.

Then invoke the debugging-methodology skill and apply the Four Phases (root cause investigation → pattern analysis → hypothesis and testing → implementation). When the cause is unclear, use the Structured Hypothesis Investigation section to enumerate and test 3-5 hypotheses; independent hypotheses may be investigated in parallel via subagents.

Once the root cause is found, you may delegate the durable regression test to the test-writer agent. Verify it fails before the fix and passes after.