Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →This skill should be used only when the user explicitly asks to use `$ralph-specum-feedback`, or explicitly asks Ralph Specum in Codex to draft or submit feedback.
.claude/skills/tzachbon-ralph-specum-feedback/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 21% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 35% | 0% |
| case-04 | ✗→✓ | ▲ Improved | -23% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -43% | 0% |
| case-18 | ✗→✓ | ▲ Improved | -23% | 0% |
Use this to capture product feedback or bug reports for Ralph Specum.
gh is available and the user wants submission, create a GitHub issue.gh is unavailable or the user only wants a draft, produce a ready-to-paste issue body and the repository issue URL.Keep the report concrete. Include expected behavior, actual behavior, reproduction steps, and any relevant state files or logs.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 9,449 | 9,840 | +4% | 1 | 1 | 0% | 1,430 | 1,725 | +21% | 0 | 0 | — |
case-02 | fail→pass | 8,665 | 10,243 | +18% | 1 | 1 | 0% | 1,252 | 1,693 | +35% | 0 | 0 | — |
case-03 | pass→pass | 9,826 | 6,054 | -38% | 1 | 1 | 0% | 1,480 | 1,051 | -29% | 0 | 0 | — |
case-04 | fail→pass | 9,468 | 6,604 | -30% | 1 | 1 | 0% | 1,493 | 1,145 | -23% | 0 | 0 | — |
case-05 | pass→pass | 9,261 | 7,210 | -22% | 1 | 1 | 0% | 1,317 | 1,219 | -7% | 0 | 0 | — |
case-06 | pass→fail | 10,655 | 7,206 | -32% | 1 | 1 | 0% | 1,553 | 1,202 | -23% | 0 | 0 | — |
case-07 | fail→pass | 14,288 | 7,148 | -50% | 1 | 1 | 0% | 2,145 | 1,218 | -43% | 0 | 0 | — |
case-12 | pass→pass | 9,100 | 5,754 | -37% | 1 | 1 | 0% | 1,417 | 1,108 | -22% | 0 | 0 | — |
case-08 | pass→pass | 9,108 | 9,026 | -1% | 1 | 1 | 0% | 1,593 | 1,626 | +2% | 0 | 0 | — |
case-09 | pass→pass | 9,218 | 5,847 | -37% | 1 | 1 | 0% | 1,417 | 1,017 | -28% | 0 | 0 | — |
case-10 | pass→pass | 8,665 | 6,799 | -22% | 1 | 1 | 0% | 1,310 | 1,114 | -15% | 0 | 0 | — |
case-11 | pass→pass | 12,044 | 8,676 | -28% | 1 | 1 | 0% | 1,850 | 1,419 | -23% | 0 | 0 | — |
case-13 | pass→pass | 12,811 | 6,972 | -46% | 1 | 1 | 0% | 1,945 | 1,183 | -39% | 0 | 0 | — |
case-14 | pass→pass | 6,611 | 10,733 | +62% | 1 | 1 | 0% | 914 | 1,028 | +12% | 0 | 0 | — |
case-15 | pass→pass | 10,126 | 7,208 | -29% | 1 | 1 | 0% | 1,485 | 1,163 | -22% | 0 | 0 | — |
case-16 | pass→pass | 11,451 | 7,069 | -38% | 1 | 1 | 0% | 1,737 | 1,004 | -42% | 0 | 0 | — |
case-17 | pass→pass | 13,456 | 6,252 | -54% | 1 | 1 | 0% | 2,017 | 1,113 | -45% | 0 | 0 | — |
case-18 | fail→pass | 10,897 | 7,239 | -34% | 1 | 1 | 0% | 1,571 | 1,203 | -23% | 0 | 0 | — |
case-19 | pass→pass | 7,855 | 5,474 | -30% | 1 | 1 | 0% | 1,187 | 1,109 | -7% | 0 | 0 | — |
case-20 | pass→fail | 13,023 | 9,720 | -25% | 1 | 1 | 0% | 2,257 | 1,630 | -28% | 0 | 0 | — |
case-21 | fail→fail | 5,516 | 3,154 | -43% | 1 | 1 | 0% | 746 | 557 | -25% | 0 | 0 | — |
case-22 | pass→pass | 9,013 | 4,683 | -48% | 1 | 1 | 0% | 1,439 | 891 | -38% | 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. 22 cases were attempted. The headline lift of +14 percentage points is the difference between those two pass rates over the 22 comparable cases. 2 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.