Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Adversaries may use voice communications to elicit sensitive information that can be used during targeting.
.claude/skills/cyberstrikeus-t1598-004-spearphishing-voice/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-04 | ✗→✓ | ▲ Improved | 57% | 0% |
| case-01 | ✗→✓ | ▲ Improved | 65% | 0% |
| case-09 | ✗→✓ | ▲ Improved | -22% | 0% |
| case-14 | ✗→✓ | ▲ Improved | -47% | 0% |
| case-15 | ✗→✓ | ▲ Improved | -58% | 0% |
> Sub-technique of: T1598
Adversaries may use voice communications to elicit sensitive information that can be used during targeting. Spearphishing for information is an attempt to trick targets into divulging information, frequently credentials or other actionable information. Spearphishing for information frequently involves social engineering techniques, such as posing as a source with a reason to collect information (ex: Impersonation) and/or creating a sense of urgency or alarm for the recipient.
All forms of phishing are electronically delivered social engineering. In this scenario, adversaries use phone calls to elicit sensitive information from victims. Known as voice phishing (or "vishing"), these communications can be manually executed by adversaries, hired call centers, or even automated via robocalls. Voice phishers may spoof their phone number while also posing as a trusted entity, such as a business partner or technical support staff.
Victims may also receive phishing messages that direct them to call a phone number ("callback phishing") where the adversary attempts to collect confidential information.
Adversaries may also use information from previous reconnaissance efforts (ex: Search Open Websites/Domains or Search Victim-Owned Websites) to tailor pretexts to be even more persuasive and believable for the victim.
Platforms: PRE
> Note: No Atomic Red Team tests available for this technique. See Atomic Red Team GitHub for updates.
Users can be trained to identify and report social engineering techniques and spearphishing attempts, while also being suspicious of and verifying the identify of callers.
| Finding | Severity | Impact | | ---------------------------------------- | -------- | -------------- | | Spearphishing Voice technique applicable | High | Reconnaissance |
| 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→fail | 33,166 | 35,685 | +8% | 1 | 1 | 0% | 4,215 | 4,853 | +15% | 0 | 0 | — |
case-03 | pass→pass | 16,074 | 15,390 | -4% | 1 | 1 | 0% | 1,904 | 2,612 | +37% | 0 | 0 | — |
case-04 | fail→pass | 16,653 | 14,668 | -12% | 1 | 1 | 0% | 1,555 | 2,447 | +57% | 0 | 0 | — |
case-05 | pass→pass | 16,761 | 15,745 | -6% | 1 | 1 | 0% | 1,577 | 2,565 | +63% | 0 | 0 | — |
case-06 | pass→pass | 15,733 | 17,199 | +9% | 1 | 1 | 0% | 1,725 | 2,465 | +43% | 0 | 0 | — |
case-11 | pass→pass | 7,270 | 8,418 | +16% | 1 | 1 | 0% | 1,281 | 1,215 | -5% | 0 | 0 | — |
case-01 | fail→pass | 24,076 | 30,041 | +25% | 1 | 1 | 0% | 2,748 | 4,536 | +65% | 0 | 0 | — |
case-07 | pass→pass | 13,307 | 8,375 | -37% | 1 | 1 | 0% | 1,089 | 1,360 | +25% | 0 | 0 | — |
case-08 | pass→pass | 18,234 | 18,313 | +0% | 1 | 1 | 0% | 1,956 | 1,974 | +1% | 0 | 0 | — |
case-09 | fail→pass | 19,002 | 6,536 | -66% | 1 | 1 | 0% | 2,162 | 1,685 | -22% | 0 | 0 | — |
case-10 | pass→pass | 14,244 | 3,913 | -73% | 1 | 1 | 0% | 1,786 | 1,034 | -42% | 0 | 0 | — |
case-12 | pass→pass | 8,945 | 8,520 | -5% | 1 | 1 | 0% | 1,630 | 1,251 | -23% | 0 | 0 | — |
case-13 | pass→pass | 16,048 | 9,176 | -43% | 1 | 1 | 0% | 1,515 | 1,343 | -11% | 0 | 0 | — |
case-14 | fail→pass | 23,126 | 11,386 | -51% | 1 | 1 | 0% | 3,075 | 1,641 | -47% | 0 | 0 | — |
case-15 | fail→pass | 22,560 | 7,746 | -66% | 1 | 1 | 0% | 2,819 | 1,179 | -58% | 0 | 0 | — |
case-16 | pass→pass | 37,547 | 8,183 | -78% | 1 | 1 | 0% | 4,898 | 1,311 | -73% | 0 | 0 | — |
case-17 | pass→pass | 12,390 | 2,741 | -78% | 1 | 1 | 0% | 2,360 | 1,081 | -54% | 0 | 0 | — |
case-18 | pass→pass | 10,651 | 7,568 | -29% | 1 | 1 | 0% | 933 | 1,239 | +33% | 0 | 0 | — |
case-19 | pass→pass | 6,360 | 3,820 | -40% | 1 | 1 | 0% | 1,199 | 1,091 | -9% | 0 | 0 | — |
case-20 | fail→pass | 14,195 | 8,699 | -39% | 1 | 1 | 0% | 1,296 | 1,127 | -13% | 0 | 0 | — |
case-21 | pass→pass | 11,983 | 17,860 | +49% | 1 | 1 | 0% | 1,676 | 2,785 | +66% | 0 | 0 | — |
case-22 | fail→fail | 9,262 | 2,043 | -78% | 1 | 1 | 0% | 755 | 1,183 | +57% | 0 | 0 | — |
case-23 | pass→pass | 18,555 | 15,858 | -15% | 1 | 1 | 0% | 1,710 | 2,126 | +24% | 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 +26 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.