Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Adversaries may modify systems in order to manipulate the data as it is accessed and displayed to an end user, thus threatening the integrity of the data.
.claude/skills/cyberstrikeus-t1565-003-runtime-data-manipulation/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-17 | ✗→✓ | ▲ Improved | -20% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 64% | 0% |
| case-04 | ✗→✓ | ▲ Improved | -26% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -47% | 0% |
| case-09 | ✗→✓ | ▲ Improved | -8% | 0% |
> Sub-technique of: T1565
Adversaries may modify systems in order to manipulate the data as it is accessed and displayed to an end user, thus threatening the integrity of the data. By manipulating runtime data, adversaries may attempt to affect a business process, organizational understanding, and decision making.
Adversaries may alter application binaries used to display data in order to cause runtime manipulations. Adversaries may also conduct Change Default File Association and Masquerading to cause a similar effect. The type of modification and the impact it will have depends on the target application and process as well as the goals and objectives of the adversary. For complex systems, an adversary would likely need special expertise and possibly access to specialized software related to the system that would typically be gained through a prolonged information gathering campaign in order to have the desired impact.
Platforms: Linux, macOS, Windows
> Note: No Atomic Red Team tests available for this technique. See Atomic Red Team GitHub for updates.
Identify critical business and system processes that may be targeted by adversaries and work to isolate and secure those systems against unauthorized access and tampering.
Prevent critical business and system processes from being replaced, overwritten, or reconfigured to load potentially malicious code.
| Finding | Severity | Impact | | ---------------------------------------------- | -------- | ------ | | Runtime Data Manipulation technique applicable | Low | Impact |
| CWE ID | Title | | ------- | --------------------------------- | | CWE-400 | Uncontrolled Resource Consumption |
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-17 | fail→pass | 15,980 | 12,498 | -22% | 1 | 1 | 0% | 2,499 | 2,002 | -20% | 0 | 0 | — |
case-01 | fail→fail | 29,894 | 39,921 | +34% | 1 | 1 | 0% | 5,037 | 6,425 | +28% | 0 | 0 | — |
case-02 | fail→fail | 33,236 | 35,906 | +8% | 1 | 1 | 0% | 4,348 | 5,692 | +31% | 0 | 0 | — |
case-03 | fail→pass | 36,335 | 48,933 | +35% | 1 | 1 | 0% | 4,758 | 7,816 | +64% | 0 | 0 | — |
case-04 | fail→pass | 20,177 | 9,859 | -51% | 1 | 1 | 0% | 3,096 | 2,291 | -26% | 0 | 0 | — |
case-05 | fail→pass | 20,956 | 9,392 | -55% | 1 | 1 | 0% | 2,738 | 1,463 | -47% | 0 | 0 | — |
case-23 | pass→pass | 24,169 | 11,288 | -53% | 1 | 1 | 0% | 2,852 | 1,983 | -30% | 0 | 0 | — |
case-06 | pass→pass | 19,939 | 12,110 | -39% | 1 | 1 | 0% | 2,025 | 1,928 | -5% | 0 | 0 | — |
case-07 | fail→fail | 24,589 | 22,147 | -10% | 1 | 1 | 0% | 2,638 | 3,267 | +24% | 0 | 0 | — |
case-08 | fail→fail | 21,815 | 25,100 | +15% | 1 | 1 | 0% | 3,193 | 4,175 | +31% | 0 | 0 | — |
case-09 | fail→pass | 20,154 | 13,674 | -32% | 1 | 1 | 0% | 2,362 | 2,177 | -8% | 0 | 0 | — |
case-10 | pass→pass | 4,765 | 4,247 | -11% | 1 | 1 | 0% | 837 | 1,546 | +85% | 0 | 0 | — |
case-11 | pass→pass | 7,807 | 12,324 | +58% | 1 | 1 | 0% | 1,225 | 1,919 | +57% | 0 | 0 | — |
case-12 | pass→pass | 14,284 | 2,607 | -82% | 1 | 1 | 0% | 2,420 | 1,260 | -48% | 0 | 0 | — |
case-13 | pass→pass | 28,940 | 25,888 | -11% | 1 | 1 | 0% | 3,604 | 4,143 | +15% | 0 | 0 | — |
case-14 | fail→pass | 18,138 | 5,422 | -70% | 1 | 1 | 0% | 2,150 | 1,625 | -24% | 0 | 0 | — |
case-15 | pass→pass | 15,929 | 19,849 | +25% | 1 | 1 | 0% | 2,475 | 2,870 | +16% | 0 | 0 | — |
case-16 | fail→pass | 24,261 | 5,219 | -78% | 1 | 1 | 0% | 2,790 | 1,608 | -42% | 0 | 0 | — |
case-18 | pass→pass | 20,779 | 12,741 | -39% | 1 | 1 | 0% | 2,568 | 2,915 | +14% | 0 | 0 | — |
case-19 | pass→pass | 49,983 | 41,451 | -17% | 1 | 1 | 0% | 5,818 | 5,577 | -4% | 0 | 0 | — |
case-20 | pass→pass | 24,095 | 29,935 | +24% | 1 | 1 | 0% | 2,973 | 3,639 | +22% | 0 | 0 | — |
case-21 | pass→pass | 24,214 | 34,182 | +41% | 1 | 1 | 0% | 4,182 | 5,088 | +22% | 0 | 0 | — |
case-22 | fail→pass | 18,097 | 4,675 | -74% | 1 | 1 | 0% | 2,094 | 1,598 | -24% | 0 | 0 | — |
case-24 | fail→pass | 15,327 | 8,354 | -45% | 1 | 1 | 0% | 1,730 | 1,354 | -22% | 0 | 0 | — |
case-25 | fail→pass | 18,381 | 25,356 | +38% | 1 | 1 | 0% | 2,970 | 3,844 | +29% | 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. 25 cases were attempted. The headline lift of +40 percentage points is the difference between those two pass rates over the 25 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.