Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →A malicious application can inject input to the user interface to mimic user interaction through the abuse of Android's accessibility APIs.
.claude/skills/cyberstrikeus-t1516-input-injection/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-07 | ✗→✓ | ▲ Improved | -59% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 37% | 0% |
| case-04 | ✗→✓ | ▲ Improved | 28% | 0% |
| case-10 | ✗→✓ | ▲ Improved | -59% | 0% |
| case-12 | ✗→✓ | ▲ Improved | -67% | 0% |
A malicious application can inject input to the user interface to mimic user interaction through the abuse of Android's accessibility APIs.
Input Injection can be achieved using any of the following methods:
GLOBAL_ACTION_BACK (programatically mimicking a physical back button press), to trigger actions on behalf of the user.Platforms: Android
Determine if the target mobile environment is susceptible to Input Injection by examining the target platforms (Android).
Review whether mitigations for T1516 are in place. If defenses are absent or misconfigured, this technique may be exploitable.
Users should be warned against granting access to accessibility features, and to carefully scrutinize applications that request this dangerous permission.
An EMM/MDM can use the Android DevicePolicyManager.setPermittedAccessibilityServices method to set an explicit list of applications that are allowed to use Android's accessibility features.
| Finding | Severity | Impact | | ------------------------------------ | -------- | --------------- | | Input Injection technique applicable | High | Defense Evasion |
| CWE ID | Title | | ------- | ---------------------------- | | CWE-693 | Protection Mechanism Failure |
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-07 | fail→pass | 20,381 | 9,952 | -51% | 1 | 1 | 0% | 3,366 | 1,395 | -59% | 0 | 0 | — |
case-01 | fail→fail | 23,887 | 27,846 | +17% | 1 | 1 | 0% | 2,906 | 3,794 | +31% | 0 | 0 | — |
case-02 | fail→pass | 28,481 | 38,298 | +34% | 1 | 1 | 0% | 3,988 | 5,475 | +37% | 0 | 0 | — |
case-03 | fail→fail | 25,297 | 32,668 | +29% | 1 | 1 | 0% | 3,020 | 4,588 | +52% | 0 | 0 | — |
case-04 | fail→pass | 26,934 | 28,466 | +6% | 1 | 1 | 0% | 3,256 | 4,180 | +28% | 0 | 0 | — |
case-05 | pass→pass | 18,034 | 19,771 | +10% | 1 | 1 | 0% | 2,949 | 2,904 | -2% | 0 | 0 | — |
case-06 | pass→pass | 20,261 | 13,760 | -32% | 1 | 1 | 0% | 2,993 | 2,975 | -1% | 0 | 0 | — |
case-08 | fail→fail | 12,620 | 5,021 | -60% | 1 | 1 | 0% | 1,260 | 1,350 | +7% | 0 | 0 | — |
case-09 | pass→pass | 43,293 | 6,556 | -85% | 1 | 1 | 0% | 2,676 | 1,149 | -57% | 0 | 0 | — |
case-10 | fail→pass | 19,834 | 9,161 | -54% | 1 | 1 | 0% | 2,647 | 1,087 | -59% | 0 | 0 | — |
case-11 | pass→pass | 7,108 | 4,481 | -37% | 1 | 1 | 0% | 824 | 962 | +17% | 0 | 0 | — |
case-12 | fail→pass | 28,146 | 10,248 | -64% | 1 | 1 | 0% | 3,785 | 1,260 | -67% | 0 | 0 | — |
case-13 | pass→pass | 22,989 | 8,115 | -65% | 1 | 1 | 0% | 2,258 | 980 | -57% | 0 | 0 | — |
case-14 | fail→pass | 23,568 | 10,886 | -54% | 1 | 1 | 0% | 2,841 | 1,451 | -49% | 0 | 0 | — |
case-15 | pass→pass | 8,961 | 9,420 | +5% | 1 | 1 | 0% | 1,388 | 919 | -34% | 0 | 0 | — |
case-16 | pass→pass | 10,917 | 9,884 | -9% | 1 | 1 | 0% | 1,399 | 1,839 | +31% | 0 | 0 | — |
case-17 | pass→pass | 15,575 | 9,750 | -37% | 1 | 1 | 0% | 1,483 | 1,298 | -12% | 0 | 0 | — |
case-18 | pass→pass | 17,721 | 9,236 | -48% | 1 | 1 | 0% | 1,625 | 1,175 | -28% | 0 | 0 | — |
case-19 | pass→pass | 23,212 | 17,935 | -23% | 1 | 1 | 0% | 2,757 | 2,625 | -5% | 0 | 0 | — |
case-20 | fail→pass | 19,027 | 8,470 | -55% | 1 | 1 | 0% | 1,732 | 1,075 | -38% | 0 | 0 | — |
case-21 | fail→pass | 20,416 | 34,704 | +70% | 1 | 1 | 0% | 1,850 | 919 | -50% | 0 | 0 | — |
case-22 | pass→pass | 21,368 | 47,627 | +123% | 1 | 1 | 0% | 2,428 | 1,322 | -46% | 0 | 0 | — |
case-23 | fail→pass | 33,157 | 12,342 | -63% | 1 | 1 | 0% | 2,138 | 917 | -57% | 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 +39 percentage points is the difference between those two pass rates over the 23 comparable cases.
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.