Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Expert in secure mobile coding practices specializing in input validation, WebView security, and mobile-specific security patterns. Use PROACTIVELY for mobile security implementations or mobile security code reviews.
.claude/skills/dokhacgiakhoa-mobile-security-coder/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-09 | ✗→✓ | ▲ Improved | 26% | 0% |
| case-10 | ✗→✓ | ▲ Improved | 43% | 0% |
| case-12 | ✓→✗ | ▼ Worse | 54% | 0% |
| case-02 | ✓→✓ | ＝ Same ✓ | 10% | 0% |
| case-03 | ✓→✓ | ＝ Same ✓ | 30% | 0% |
resources/implementation-playbook.md.You are a mobile security coding expert specializing in secure mobile development practices, mobile-specific vulnerabilities, and secure mobile architecture patterns.
Expert mobile security developer with comprehensive knowledge of mobile security practices, platform-specific vulnerabilities, and secure mobile application development. Masters input validation, WebView security, secure data storage, and mobile authentication patterns. Specializes in building security-first mobile applications that protect sensitive data and resist mobile-specific attack vectors.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 17,902 | 25,580 | +43% | 1 | 1 | 0% | 3,348 | 4,323 | +29% | 0 | 0 | — |
case-02 | pass→pass | 18,725 | 18,416 | -2% | 1 | 1 | 0% | 3,980 | 4,372 | +10% | 0 | 0 | — |
case-03 | pass→pass | 18,346 | 16,695 | -9% | 1 | 1 | 0% | 2,843 | 3,699 | +30% | 0 | 0 | — |
case-04 | fail→fail | 14,161 | 13,499 | -5% | 1 | 1 | 0% | 2,332 | 3,203 | +37% | 0 | 0 | — |
case-05 | pass→pass | 22,871 | 28,439 | +24% | 1 | 1 | 0% | 3,412 | 5,034 | +48% | 0 | 0 | — |
case-06 | fail→fail | 18,504 | 17,955 | -3% | 1 | 1 | 0% | 3,289 | 4,362 | +33% | 0 | 0 | — |
case-07 | pass→pass | 14,698 | 15,978 | +9% | 1 | 1 | 0% | 2,549 | 3,460 | +36% | 0 | 0 | — |
case-08 | pass→pass | 14,766 | 12,393 | -16% | 1 | 1 | 0% | 2,148 | 2,702 | +26% | 0 | 0 | — |
case-09 | fail→pass | 17,160 | 21,104 | +23% | 1 | 1 | 0% | 2,911 | 3,656 | +26% | 0 | 0 | — |
case-10 | fail→pass | 19,689 | 24,326 | +24% | 1 | 1 | 0% | 3,636 | 5,190 | +43% | 0 | 0 | — |
case-19 | pass→pass | 8,892 | 12,929 | +45% | 1 | 1 | 0% | 1,743 | 2,393 | +37% | 0 | 0 | — |
case-11 | pass→pass | 12,276 | 12,167 | -1% | 1 | 1 | 0% | 1,848 | 2,832 | +53% | 0 | 0 | — |
case-12 | pass→fail | 14,444 | 17,599 | +22% | 1 | 1 | 0% | 2,420 | 3,734 | +54% | 0 | 0 | — |
case-13 | pass→pass | 19,501 | 21,035 | +8% | 1 | 1 | 0% | 2,690 | 3,795 | +41% | 0 | 0 | — |
case-14 | pass→pass | 11,199 | 15,496 | +38% | 1 | 1 | 0% | 1,821 | 3,698 | +103% | 0 | 0 | — |
case-15 | pass→pass | 19,162 | 18,354 | -4% | 1 | 1 | 0% | 2,743 | 3,727 | +36% | 0 | 0 | — |
case-16 | fail→fail | 22,309 | 22,660 | +2% | 1 | 1 | 0% | 3,271 | 5,073 | +55% | 0 | 0 | — |
case-17 | pass→pass | 10,988 | 10,404 | -5% | 1 | 1 | 0% | 1,624 | 2,102 | +29% | 0 | 0 | — |
case-18 | pass→pass | 20,963 | 21,410 | +2% | 1 | 1 | 0% | 3,210 | 3,976 | +24% | 0 | 0 | — |
case-20 | fail→fail | 25,573 | 39,249 | +53% | 1 | 1 | 0% | 4,694 | 7,409 | +58% | 0 | 0 | — |
case-21 | fail→fail | 32,183 | 32,227 | +0% | 1 | 1 | 0% | 4,887 | 6,331 | +30% | 0 | 0 | — |
case-22 | pass→pass | 50,394 | 43,308 | -14% | 1 | 1 | 0% | 8,216 | 6,638 | -19% | 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 +5 percentage points is the difference between those two pass rates over the 22 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.
Other measured skills in the registry, with their headline benchmark lift.