Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Adversaries may use an existing, legitimate external Web service to exfiltrate data rather than their primary command and control channel.
.claude/skills/cyberstrikeus-t1567-exfiltration-over-web-service/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-08 | ✗→✓ | ▲ Improved | -13% | 0% |
| case-10 | ✗→✓ | ▲ Improved | -7% | 0% |
| case-14 | ✗→✓ | ▲ Improved | -27% | 0% |
| case-16 | ✗→✓ | ▲ Improved | -18% | 0% |
| case-24 | ✗→✓ | ▲ Improved | 45% | 0% |
Adversaries may use an existing, legitimate external Web service to exfiltrate data rather than their primary command and control channel. Popular Web services acting as an exfiltration mechanism may give a significant amount of cover due to the likelihood that hosts within a network are already communicating with them prior to compromise. Firewall rules may also already exist to permit traffic to these services.
Web service providers also commonly use SSL/TLS encryption, giving adversaries an added level of protection.
Platforms: ESXi, Linux, macOS, Office Suite, SaaS, Windows
> Note: No Atomic Red Team tests available for this technique. See Atomic Red Team GitHub for updates.
Web proxies can be used to enforce an external network communication policy that prevents use of unauthorized external services.
Data loss prevention can be detect and block sensitive data being uploaded to web services via web browsers.
| Finding | Severity | Impact | | -------------------------------------------------- | -------- | ------------ | | Exfiltration Over Web Service technique applicable | Low | Exfiltration |
| CWE ID | Title | | ------- | --------------------------------- | | CWE-200 | Exposure of Sensitive Information |
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 46,229 | 40,413 | -13% | 1 | 1 | 0% | 6,300 | 6,532 | +4% | 0 | 0 | — |
case-02 | fail→fail | 23,592 | 26,959 | +14% | 1 | 1 | 0% | 3,708 | 5,045 | +36% | 0 | 0 | — |
case-03 | fail→fail | 20,747 | 46,831 | +126% | 1 | 1 | 0% | 3,132 | 6,414 | +105% | 0 | 0 | — |
case-04 | pass→pass | 19,698 | 19,419 | -1% | 1 | 1 | 0% | 2,319 | 2,927 | +26% | 0 | 0 | — |
case-05 | pass→pass | 22,206 | 30,809 | +39% | 1 | 1 | 0% | 3,481 | 6,282 | +80% | 0 | 0 | — |
case-06 | fail→fail | 18,811 | 26,457 | +41% | 1 | 1 | 0% | 3,280 | 4,739 | +44% | 0 | 0 | — |
case-07 | fail→fail | 14,245 | 20,996 | +47% | 1 | 1 | 0% | 1,538 | 3,578 | +133% | 0 | 0 | — |
case-08 | fail→pass | 12,608 | 2,762 | -78% | 1 | 1 | 0% | 1,190 | 1,039 | -13% | 0 | 0 | — |
case-09 | pass→pass | 14,101 | 2,669 | -81% | 1 | 1 | 0% | 1,432 | 899 | -37% | 0 | 0 | — |
case-10 | fail→pass | 13,000 | 14,221 | +9% | 1 | 1 | 0% | 1,975 | 1,845 | -7% | 0 | 0 | — |
case-11 | pass→pass | 14,458 | 7,205 | -50% | 1 | 1 | 0% | 1,761 | 892 | -49% | 0 | 0 | — |
case-12 | pass→pass | 19,147 | 13,219 | -31% | 1 | 1 | 0% | 2,266 | 1,967 | -13% | 0 | 0 | — |
case-13 | pass→pass | 11,200 | 9,892 | -12% | 1 | 1 | 0% | 1,598 | 1,268 | -21% | 0 | 0 | — |
case-14 | fail→pass | 12,018 | 7,133 | -41% | 1 | 1 | 0% | 1,389 | 1,009 | -27% | 0 | 0 | — |
case-15 | pass→pass | 19,676 | 6,816 | -65% | 1 | 1 | 0% | 2,595 | 914 | -65% | 0 | 0 | — |
case-16 | fail→pass | 11,292 | 6,443 | -43% | 1 | 1 | 0% | 1,071 | 879 | -18% | 0 | 0 | — |
case-17 | pass→pass | 11,051 | 2,243 | -80% | 1 | 1 | 0% | 1,058 | 948 | -10% | 0 | 0 | — |
case-18 | pass→pass | 1,963 | 6,620 | +237% | 1 | 1 | 0% | 371 | 906 | +144% | 0 | 0 | — |
case-19 | pass→pass | 19,318 | 4,833 | -75% | 1 | 1 | 0% | 2,017 | 1,047 | -48% | 0 | 0 | — |
case-20 | pass→pass | 15,888 | 8,118 | -49% | 1 | 1 | 0% | 1,834 | 1,007 | -45% | 0 | 0 | — |
case-21 | pass→pass | 19,683 | 17,331 | -12% | 1 | 1 | 0% | 2,327 | 3,267 | +40% | 0 | 0 | — |
case-22 | pass→pass | 20,534 | 7,429 | -64% | 1 | 1 | 0% | 1,835 | 903 | -51% | 0 | 0 | — |
case-23 | pass→pass | 16,750 | 6,534 | -61% | 1 | 1 | 0% | 1,930 | 882 | -54% | 0 | 0 | — |
case-24 | fail→pass | 8,870 | 6,685 | -25% | 1 | 1 | 0% | 627 | 908 | +45% | 0 | 0 | — |
case-25 | fail→fail | 18,635 | 16,323 | -12% | 1 | 1 | 0% | 2,478 | 2,594 | +5% | 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 +20 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.