Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Create beautiful visual art in .png and .pdf documents using design philosophy. You should use this skill when the user asks to create a poster, piece of art, design, or other static piece. Create original visual designs, never copying existing artists' work to avoid copyright violations.
.claude/skills/dokhacgiakhoa-canvas-design/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-06 | ✗→✓ | ▲ Improved | 72% | 0% |
| case-09 | ✗→✓ | ▲ Improved | -11% | 0% |
| case-15 | ✗→✓ | ▲ Improved | 180% | 0% |
| case-02 | ✓→✗ | ▼ Worse | 22% | 0% |
| case-14 | ✓→✗ | ▼ Worse | 26% | 0% |
These are instructions for creating design philosophies - aesthetic movements that are then EXPRESSED VISUALLY. Output only .md files, .pdf files, and .png files.
Complete this in two steps:
First, undertake this task:
To begin, create a VISUAL PHILOSOPHY (not layouts or templates) that will be interpreted through:
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 11,239 | 42,888 | +282% | 1 | 1 | 0% | 1,703 | 6,117 | +259% | 0 | 0 | — |
case-02 | pass→fail | 33,044 | 41,895 | +27% | 1 | 1 | 0% | 6,955 | 8,488 | +22% | 0 | 0 | — |
case-03 | fail→fail | 17,188 | 22,339 | +30% | 1 | 1 | 0% | 2,455 | 3,800 | +55% | 0 | 0 | — |
case-04 | fail→fail | 39,154 | 31,568 | -19% | 1 | 1 | 0% | 6,111 | 6,655 | +9% | 0 | 0 | — |
case-05 | fail→fail | 43,567 | 44,080 | +1% | 1 | 1 | 0% | 8,239 | 7,802 | -5% | 0 | 0 | — |
case-06 | fail→pass | 18,461 | 26,639 | +44% | 1 | 1 | 0% | 2,811 | 4,832 | +72% | 0 | 0 | — |
case-07 | fail→fail | 18,686 | 43,883 | +135% | 1 | 1 | 0% | 3,283 | 8,239 | +151% | 0 | 0 | — |
case-08 | pass→pass | 34,558 | 38,716 | +12% | 1 | 1 | 0% | 7,064 | 8,481 | +20% | 0 | 0 | — |
case-09 | fail→pass | 41,587 | 47,873 | +15% | 1 | 1 | 0% | 8,235 | 7,319 | -11% | 0 | 0 | — |
case-10 | fail→fail | 37,309 | 54,957 | +47% | 1 | 1 | 0% | 6,982 | 8,665 | +24% | 0 | 0 | — |
case-11 | fail→fail | 30,766 | 78,640 | +156% | 1 | 1 | 0% | 5,101 | 6,435 | +26% | 0 | 0 | — |
case-12 | pass→pass | 20,378 | 32,647 | +60% | 1 | 1 | 0% | 2,770 | 5,132 | +85% | 0 | 0 | — |
case-13 | fail→fail | 55,516 | 42,361 | -24% | 1 | 1 | 0% | 8,233 | 8,482 | +3% | 0 | 0 | — |
case-14 | pass→fail | 51,916 | 78,712 | +52% | 1 | 1 | 0% | 8,239 | 10,370 | +26% | 0 | 0 | — |
case-15 | fail→pass | 17,243 | 40,679 | +136% | 1 | 1 | 0% | 2,225 | 6,228 | +180% | 0 | 0 | — |
case-16 | fail→fail | 53,504 | 106,272 | +99% | 1 | 1 | 0% | 8,228 | 8,477 | +3% | 0 | 0 | — |
case-17 | fail→fail | 39,574 | 65,765 | +66% | 1 | 1 | 0% | 6,328 | 6,589 | +4% | 0 | 0 | — |
case-18 | fail→fail | 97,694 | 42,683 | -56% | 1 | 1 | 0% | 6,459 | 8,486 | +31% | 0 | 0 | — |
case-19 | pass→pass | 42,806 | 42,241 | -1% | 1 | 1 | 0% | 6,230 | 7,859 | +26% | 0 | 0 | — |
case-20 | fail→fail | 9,133 | 39,235 | +330% | 1 | 1 | 0% | 1,277 | 7,421 | +481% | 0 | 0 | — |
case-21 | fail→fail | 49,880 | 41,152 | -17% | 1 | 1 | 0% | 7,671 | 8,483 | +11% | 0 | 0 | — |
case-22 | pass→fail | 12,756 | 82,202 | +544% | 1 | 1 | 0% | 2,925 | 8,399 | +187% | 0 | 0 | — |
case-23 | pass→fail | 31,031 | 62,762 | +102% | 1 | 1 | 0% | 5,045 | 8,471 | +68% | 0 | 0 | — |
case-24 | pass→fail | 33,253 | 23,399 | -30% | 1 | 1 | 0% | 8,104 | 3,784 | -53% | 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. 24 cases were attempted, and 23 counted toward the lift figure. The other 1 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 -8 percentage points is the difference between those two pass rates over the 23 comparable cases. 6 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.