Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Create a new SKILL.md file based on a guided conversation about a task or workflow.
.claude/skills/abdullahkhawer-create-skill/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -28% | 0% |
| case-02 | ✗→✓ | ▲ Improved | -64% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -73% | 0% |
| case-04 | ✗→✓ | ▲ Improved | -49% | 0% |
| case-05 | ✗→✓ | ▲ Improved | 44% | 0% |
Run this skill when asked to create a new skill from a conversation or description of a task.
Ask the user only these two questions:
1. What specific task or problem do you want this skill to solve?
2. What are the specific instructions and guidelines for this task?Wait for the user's answers before proceeding.
Derive a short, lowercase, hyphenated skill name from the task description (e.g. deploy-lambda, check-postgres-health).
Ask the user which AI agent they are using, then use the corresponding path:
| Agent | Skill path | |---|---| | Claude Code | .claude/skills/<skill-name>/SKILL.md | | GitHub Copilot | .github/skills/<skill-name>/SKILL.md | | Other | skills/<skill-name>/SKILL.md |
Create the directory if it does not exist:
bashmkdir -p <path-to-skill-dir>
Use the template below, filling in all <...> placeholders based on the user's answers.
markdown--- name: <skill-name> description: <One-liner description of what this skill does.> --- # <Skill Title> Run this skill when asked to <task statement>. ## Step 1 — <First step title> <Instructions for the first step.> ## Step 2 — <Second step title> <Instructions for the second step.> ... ## Rules - <Rule 1> - <Rule 2>
After writing the file, show the user the generated SKILL.md content and the path where it was saved. Ask if any adjustments are needed.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 6,384 | 2,907 | -54% | 1 | 1 | 0% | 1,086 | 780 | -28% | 0 | 0 | — |
case-02 | fail→pass | 11,656 | 12,073 | +4% | 1 | 1 | 0% | 2,309 | 820 | -64% | 0 | 0 | — |
case-03 | fail→pass | 14,519 | 1,748 | -88% | 1 | 1 | 0% | 3,093 | 842 | -73% | 0 | 0 | — |
case-04 | fail→pass | 9,057 | 1,702 | -81% | 1 | 1 | 0% | 1,562 | 793 | -49% | 0 | 0 | — |
case-05 | fail→pass | 17,388 | 4,920 | -72% | 1 | 1 | 0% | 1,087 | 1,566 | +44% | 0 | 0 | — |
case-06 | fail→pass | 9,035 | 1,606 | -82% | 1 | 1 | 0% | 1,476 | 902 | -39% | 0 | 0 | — |
case-07 | pass→pass | 7,469 | 1,691 | -77% | 1 | 1 | 0% | 1,266 | 865 | -32% | 0 | 0 | — |
case-08 | pass→pass | 9,197 | 25,798 | +181% | 1 | 1 | 0% | 1,858 | 951 | -49% | 0 | 0 | — |
case-09 | pass→pass | 5,481 | 2,135 | -61% | 1 | 1 | 0% | 989 | 965 | -2% | 0 | 0 | — |
case-10 | pass→pass | 4,187 | 1,608 | -62% | 1 | 1 | 0% | 564 | 876 | +55% | 0 | 0 | — |
case-11 | pass→pass | 7,957 | 5,675 | -29% | 1 | 1 | 0% | 1,462 | 1,704 | +17% | 0 | 0 | — |
case-12 | pass→pass | 8,251 | 3,380 | -59% | 1 | 1 | 0% | 1,409 | 1,169 | -17% | 0 | 0 | — |
case-13 | pass→pass | 11,391 | 3,462 | -70% | 1 | 1 | 0% | 2,022 | 1,252 | -38% | 0 | 0 | — |
case-14 | pass→pass | 5,298 | 1,852 | -65% | 1 | 1 | 0% | 880 | 899 | +2% | 0 | 0 | — |
case-15 | pass→pass | 10,886 | 18,891 | +74% | 1 | 1 | 0% | 1,713 | 1,193 | -30% | 0 | 0 | — |
case-16 | pass→pass | 9,159 | 4,733 | -48% | 1 | 1 | 0% | 1,614 | 1,389 | -14% | 0 | 0 | — |
case-17 | fail→pass | 6,169 | 2,723 | -56% | 1 | 1 | 0% | 1,187 | 1,041 | -12% | 0 | 0 | — |
case-18 | fail→pass | 7,965 | 3,049 | -62% | 1 | 1 | 0% | 1,340 | 1,132 | -16% | 0 | 0 | — |
case-19 | fail→pass | 4,555 | 1,963 | -57% | 1 | 1 | 0% | 760 | 969 | +28% | 0 | 0 | — |
case-20 | pass→pass | 3,816 | 1,970 | -48% | 1 | 1 | 0% | 777 | 944 | +21% | 0 | 0 | — |
case-21 | pass→fail | 3,201 | 5,669 | +77% | 1 | 1 | 0% | 495 | 1,632 | +230% | 0 | 0 | — |
case-22 | pass→pass | 7,367 | 8,366 | +14% | 1 | 1 | 0% | 1,489 | 2,182 | +47% | 0 | 0 | — |
case-23 | pass→pass | 8,355 | 6,029 | -28% | 1 | 1 | 0% | 1,602 | 1,684 | +5% | 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. The headline lift of +35 percentage points is the difference between those two pass rates over the 23 comparable cases. 1 case got worse with the skill loaded, and it is 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.
| Model | Method | Date | Lift |
|---|---|---|---|
| gemini-3.6-flash | verified | 8/3/2026 | +27% |
Other measured skills in the registry, with their headline benchmark lift.