Agent skill · software engineering · jeremylongshore
hyperflow-trace
Hyperflow debugging. Use for bugs, test failures, runtime errors, broken builds, or "this doesn't work" reports — verbs like debug, "fix it", solve, "why is X failing", "Y is broken", or a pasted stack trace. Systematic root-cause analysis before any patch — never blind-patch symptoms.
Why this skill is useful
Provides a systematic root-cause analysis framework that the AI wouldn't reliably generate on its own.
What it needs
About 1k tokens when loaded. Last updated 2026-08-07. 2,604 stars on the source repository.
What this skill does
hyperflow-trace — root-cause phase (Antigravity single-agent) Find the root cause before changing anything. Follow the hyperflow doctrine. Steps 1. Reproduce / locate. Read the error, stack trace, or failing test. Identify the exact failing line and the observed-vs-expected behavior. 2. 5 Whys. Trace the causal chain backward — keep asking "why" until you reach the true cause, not a symptom. 3. Hypotheses. List the 2-4 most plausible causes. For each, state a cheap test (read a file, add a log, run one test) that confirms or rules it out. Test them — narrow to the real cause. 4. Confirm the root cause with evidence (a failing assertion, a value print, a reproduced path). Do not patch on a guess. 5. Fix the root cause minimally. Add or update a test that would have caught it (characterization test before behavior change). 6. Verify: re-run the failing case + the surrounding suite. Self-review the diff (L1-L3). Commit as fix(<scope>): <root cause> (conventional, lowercase). Rules Never blind-patch a symptom to make an error message disappear. No behavior change beyond the fix unless asked. If the root cause is unclear after hypothesis testing, surface what you found and what's still unknown — don't ship a speculative patch.
How to use it
Reference it in AdaL, Claude Code, Cursor or any coding agent — nothing to install:
@skills jeremylongshore/hyperflow-trace