Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →You are a frontend security specialist focusing on Cross-Site Scripting (XSS) vulnerability detection and prevention. Analyze React, Vue, Angular, and vanilla JavaScript code to identify injection poi
.claude/skills/dokhacgiakhoa-frontend-mobile-security-xss-scan/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-05 | ✗→✓ | ▲ Improved | 76% | 0% |
| case-10 | ✗→✓ | ▲ Improved | 23% | 0% |
| case-01 | ✓→✗ | ▼ Worse | -35% | 0% |
| case-18 | ✓→✗ | ▼ Worse | 4% | 0% |
| case-02 | ✓→✓ | ＝ Same ✓ | 19% | 0% |
You are a frontend security specialist focusing on Cross-Site Scripting (XSS) vulnerability detection and prevention. Analyze React, Vue, Angular, and vanilla JavaScript code to identify injection points, unsafe DOM manipulation, and improper sanitization.
The user needs comprehensive XSS vulnerability scanning for client-side code, identifying dangerous patterns like unsafe HTML manipulation, URL handling issues, and improper user input rendering. Focus on context-aware detection and framework-specific security patterns.
$ARGUMENTS
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | pass→fail | 19,215 | 28,533 | +48% | 1 | 1 | 0% | 2,933 | 1,894 | -35% | 0 | 0 | — |
case-02 | pass→pass | 17,535 | 17,799 | +2% | 1 | 1 | 0% | 2,982 | 3,560 | +19% | 0 | 0 | — |
case-03 | pass→pass | 17,273 | 18,160 | +5% | 1 | 1 | 0% | 1,693 | 2,538 | +50% | 0 | 0 | — |
case-04 | pass→pass | 8,893 | 7,888 | -11% | 1 | 1 | 0% | 1,574 | 1,753 | +11% | 0 | 0 | — |
case-05 | fail→pass | 17,277 | 25,932 | +50% | 1 | 1 | 0% | 2,909 | 5,124 | +76% | 0 | 0 | — |
case-06 | pass→pass | 15,978 | 27,406 | +72% | 1 | 1 | 0% | 2,708 | 2,939 | +9% | 0 | 0 | — |
case-07 | pass→pass | 7,723 | 11,104 | +44% | 1 | 1 | 0% | 1,335 | 2,398 | +80% | 0 | 0 | — |
case-08 | pass→pass | 11,254 | 10,968 | -3% | 1 | 1 | 0% | 2,005 | 2,329 | +16% | 0 | 0 | — |
case-09 | pass→pass | 11,871 | 13,770 | +16% | 1 | 1 | 0% | 1,960 | 2,706 | +38% | 0 | 0 | — |
case-10 | fail→pass | 15,513 | 13,833 | -11% | 1 | 1 | 0% | 2,227 | 2,742 | +23% | 0 | 0 | — |
case-11 | pass→pass | 11,360 | 13,874 | +22% | 1 | 1 | 0% | 2,014 | 2,370 | +18% | 0 | 0 | — |
case-12 | fail→fail | 12,357 | 19,056 | +54% | 1 | 1 | 0% | 2,108 | 3,413 | +62% | 0 | 0 | — |
case-13 | pass→pass | 13,036 | 16,280 | +25% | 1 | 1 | 0% | 2,122 | 3,146 | +48% | 0 | 0 | — |
case-14 | pass→pass | 10,748 | 8,227 | -23% | 1 | 1 | 0% | 1,852 | 1,750 | -6% | 0 | 0 | — |
case-15 | pass→pass | 11,576 | 10,648 | -8% | 1 | 1 | 0% | 2,023 | 2,375 | +17% | 0 | 0 | — |
case-16 | pass→pass | 6,561 | 18,972 | +189% | 1 | 1 | 0% | 1,134 | 1,972 | +74% | 0 | 0 | — |
case-17 | pass→pass | 9,048 | 8,555 | -5% | 1 | 1 | 0% | 1,579 | 1,790 | +13% | 0 | 0 | — |
case-18 | pass→fail | 15,227 | 13,540 | -11% | 1 | 1 | 0% | 2,710 | 2,807 | +4% | 0 | 0 | — |
case-19 | pass→pass | 14,150 | 12,908 | -9% | 1 | 1 | 0% | 2,290 | 2,515 | +10% | 0 | 0 | — |
case-20 | pass→pass | 12,646 | 10,237 | -19% | 1 | 1 | 0% | 2,418 | 2,292 | -5% | 0 | 0 | — |
case-21 | pass→pass | 12,288 | 13,324 | +8% | 1 | 1 | 0% | 2,291 | 2,598 | +13% | 0 | 0 | — |
case-22 | pass→pass | 14,950 | 15,832 | +6% | 1 | 1 | 0% | 3,135 | 3,675 | +17% | 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 0 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.