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.
Allowed Tools
No tools specified
Provided by Plugin
hyperflow
Fifteen specialized slash commands turn one Claude session into a structured multi-agent engineering pipeline. Thinking models orchestrate, triage, and review; worker models execute in parallel — every step is a Worker → Reviewer pair, and every non-trivial phase fans into sub-phases with their own reviewers. Auto-routing is on by default — say 'audit the diff', 'debug this test', 'large migration', or 'run a workflow' and the orchestrator routes to the right skill without the /hyperflow:* prefix. /hyperflow:workflow uses Claude Code dynamic workflows for big tasks and a portable Codex/OpenCode adapter where native workflows are unavailable; /hyperflow:spec asks the questions a senior engineer would; /hyperflow:scope decomposes into a batched task graph; /hyperflow:dispatch fans out persona-stitched workers under tiered review; /hyperflow:amplify rewrites a rough prompt into a high-quality one before you run it. 15 composable personas, 6 adaptive flow profiles, and persistent project memory compound across sessions. Works across Codex App/CLI, Claude Code, OpenCode, and Antigravity.
Installation
This skill is included in the hyperflow plugin:
/plugin install hyperflow@claude-code-plugins-plusClick to copy
Instructions
hyperflow-trace — root-cause phase (Antigravity single-agent)
Find the root cause before changing anything. Follow the hyperflow doctrine.
Steps
- Reproduce / locate. Read the error, stack trace, or failing test. Identify the exact failing line and the observed-vs-expected behavior.
- 5 Whys. Trace the causal chain backward — keep asking "why" until you reach the true cause, not a symptom.
- 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.
- Confirm the root cause with evidence (a failing assertion, a value print, a reproduced path). Do not patch on a guess.
- Fix the root cause minimally. Add or update a test that would have caught it (characterization test before behavior change).
- Verify: re-run the failing case + the surrounding suite. Self-review the diff (L1-L3). Commit as
fix((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.