Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Triage a bug or issue by exploring the codebase to find root cause, then create a GitHub issue with a TDD-based fix plan. Use when user reports a bug, wants to file an issue, mentions "triage", or wants to investigate and plan a fix for a problem.
.claude/skills/mkurman-triage-issue/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-05 | ✗→✓ | ▲ Improved | -2% | 0% |
| case-04 | ✓→✗ | ▼ Worse | 62% | 0% |
| case-10 | ✓→✗ | ▼ Worse | -22% | 0% |
| case-11 | ✓→✗ | ▼ Worse | -40% | 0% |
| case-12 | ✓→✗ | ▼ Worse | -38% | 0% |
Investigate a reported problem, find its root cause, and create a GitHub issue with a TDD fix plan. This is a mostly hands-off workflow - minimize questions to the user.
Get a brief description of the issue from the user. If they haven't provided one, ask ONE question: "What's the problem you're seeing?"
Do NOT ask follow-up questions yet. Start investigating immediately.
Use the Agent tool with subagent_type=Explore to deeply investigate the codebase. Your goal is to find:
Look at:
git log on relevant files)Based on your investigation, determine:
Create a concrete, ordered list of RED-GREEN cycles. Each cycle is one vertical slice:
Rules:
Create a GitHub issue using gh issue create with the template below. Do NOT ask the user to review before creating - just create it and share the URL.
<issue-template>
A clear description of the bug or issue, including:
Describe what you found during investigation:
Do NOT include specific file paths, line numbers, or implementation details that couple to current code layout. Describe modules, behaviors, and contracts instead. The issue should remain useful even after major refactors.
A numbered list of RED-GREEN cycles:
GREEN: Minimal change to make it pass]
GREEN: Minimal change to make it pass]
...
REFACTOR: Any cleanup needed after all tests pass]
</issue-template>
After creating the issue, print the issue URL and a one-line summary of the root cause.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-03 | fail→fail | 22,360 | 7,891 | -65% | 1 | 1 | 0% | 2,806 | 1,170 | -58% | 0 | 0 | — |
case-01 | fail→fail | 8,867 | 4,565 | -49% | 1 | 1 | 0% | 221 | 1,047 | +374% | 0 | 0 | — |
case-02 | fail→fail | 19,205 | 5,222 | -73% | 1 | 1 | 0% | 2,643 | 1,019 | -61% | 0 | 0 | — |
case-04 | pass→fail | 5,735 | 10,760 | +88% | 1 | 1 | 0% | 822 | 1,331 | +62% | 0 | 0 | — |
case-05 | fail→pass | 6,132 | 1,168 | -81% | 1 | 1 | 0% | 936 | 917 | -2% | 0 | 0 | — |
case-06 | fail→fail | 7,536 | 10,289 | +37% | 1 | 1 | 0% | 1,233 | 1,703 | +38% | 0 | 0 | — |
case-07 | fail→fail | 8,383 | 5,495 | -34% | 1 | 1 | 0% | 561 | 1,041 | +86% | 0 | 0 | — |
case-08 | pass→pass | 11,957 | 3,290 | -72% | 1 | 1 | 0% | 1,962 | 1,305 | -33% | 0 | 0 | — |
case-09 | fail→fail | 12,125 | 35,585 | +193% | 1 | 1 | 0% | 2,018 | 1,249 | -38% | 0 | 0 | — |
case-10 | pass→fail | 11,029 | 8,530 | -23% | 1 | 1 | 0% | 1,753 | 1,371 | -22% | 0 | 0 | — |
case-11 | pass→fail | 11,683 | 13,380 | +15% | 1 | 1 | 0% | 1,870 | 1,129 | -40% | 0 | 0 | — |
case-12 | pass→fail | 11,086 | 6,932 | -37% | 1 | 1 | 0% | 1,788 | 1,102 | -38% | 0 | 0 | — |
case-13 | fail→fail | 3,141 | 5,633 | +79% | 1 | 1 | 0% | 324 | 1,025 | +216% | 0 | 0 | — |
case-14 | pass→fail | 11,061 | 6,476 | -41% | 1 | 1 | 0% | 1,655 | 1,217 | -26% | 0 | 0 | — |
case-15 | pass→fail | 11,602 | 6,636 | -43% | 1 | 1 | 0% | 1,848 | 1,175 | -36% | 0 | 0 | — |
case-16 | pass→fail | 5,571 | 6,670 | +20% | 1 | 1 | 0% | 958 | 1,167 | +22% | 0 | 0 | — |
case-17 | pass→fail | 11,394 | 7,678 | -33% | 1 | 1 | 0% | 1,755 | 1,318 | -25% | 0 | 0 | — |
case-18 | fail→fail | 10,219 | 7,351 | -28% | 1 | 1 | 0% | 1,631 | 1,149 | -30% | 0 | 0 | — |
case-19 | fail→fail | 11,341 | 7,720 | -32% | 1 | 1 | 0% | 2,003 | 1,198 | -40% | 0 | 0 | — |
case-20 | pass→fail | 11,995 | 7,044 | -41% | 1 | 1 | 0% | 1,161 | 1,249 | +8% | 0 | 0 | — |
case-21 | pass→fail | 20,622 | 6,032 | -71% | 1 | 1 | 0% | 4,379 | 1,013 | -77% | 0 | 0 | — |
case-22 | fail→fail | 6,805 | 7,143 | +5% | 1 | 1 | 0% | 700 | 1,198 | +71% | 0 | 0 | — |
case-23 | pass→fail | 4,868 | 5,643 | +16% | 1 | 1 | 0% | 840 | 1,142 | +36% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 23 cases were attempted, and 3 counted toward the lift figure. The other 20 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of -43 percentage points is the difference between those two pass rates over the 3 comparable cases. 11 cases got worse with the skill loaded, and they are included in that figure.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.