Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Adversaries may send spearphishing messages via third-party services to elicit sensitive information that can be used during targeting.
.claude/skills/cyberstrikeus-t1598-001-spearphishing-service/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-06 | ✗→✓ | ▲ Improved | 29% | 0% |
| case-09 | ✗→✓ | ▲ Improved | -67% | 0% |
| case-10 | ✗→✓ | ▲ Improved | -40% | 0% |
| case-14 | ✗→✓ | ▲ Improved | -39% | 0% |
| case-15 | ✗→✓ | ▲ Improved | -47% | 0% |
> Sub-technique of: T1598
Adversaries may send spearphishing messages via third-party services 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: Establish Accounts or Compromise Accounts) and/or sending multiple, seemingly urgent messages.
All forms of spearphishing are electronically delivered social engineering targeted at a specific individual, company, or industry. In this scenario, adversaries send messages through various social media services, personal webmail, and other non-enterprise controlled services. These services are more likely to have a less-strict security policy than an enterprise. As with most kinds of spearphishing, the goal is to generate rapport with the target or get the target's interest in some way. Adversaries may create fake social media accounts and message employees for potential job opportunities. Doing so allows a plausible reason for asking about services, policies, and information about their environment. Adversaries may also use information from previous reconnaissance efforts (ex: Social Media or Search Victim-Owned Websites) to craft persuasive and believable lures.
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 social engineering techniques and spearphishing attempts.
| Finding | Severity | Impact | | ------------------------------------------ | -------- | -------------- | | Spearphishing Service 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-01 | fail→fail | 28,139 | 24,768 | -12% | 1 | 1 | 0% | 3,577 | 3,288 | -8% | 0 | 0 | — |
case-02 | fail→fail | 17,725 | 29,713 | +68% | 1 | 1 | 0% | 1,748 | 4,628 | +165% | 0 | 0 | — |
case-03 | fail→fail | 30,605 | 40,662 | +33% | 1 | 1 | 0% | 3,954 | 5,361 | +36% | 0 | 0 | — |
case-04 | pass→pass | 16,162 | 19,965 | +24% | 1 | 1 | 0% | 1,893 | 2,647 | +40% | 0 | 0 | — |
case-05 | pass→pass | 19,650 | 10,052 | -49% | 1 | 1 | 0% | 2,455 | 1,465 | -40% | 0 | 0 | — |
case-06 | fail→pass | 18,085 | 23,303 | +29% | 1 | 1 | 0% | 2,208 | 2,843 | +29% | 0 | 0 | — |
case-07 | pass→pass | 15,096 | 8,620 | -43% | 1 | 1 | 0% | 1,396 | 1,285 | -8% | 0 | 0 | — |
case-08 | pass→pass | 18,171 | 13,830 | -24% | 1 | 1 | 0% | 2,100 | 1,885 | -10% | 0 | 0 | — |
case-09 | fail→pass | 22,361 | 9,471 | -58% | 1 | 1 | 0% | 3,146 | 1,041 | -67% | 0 | 0 | — |
case-10 | fail→pass | 19,263 | 10,823 | -44% | 1 | 1 | 0% | 2,373 | 1,426 | -40% | 0 | 0 | — |
case-11 | pass→pass | 12,412 | 7,698 | -38% | 1 | 1 | 0% | 1,426 | 1,275 | -11% | 0 | 0 | — |
case-12 | pass→pass | 13,115 | 9,371 | -29% | 1 | 1 | 0% | 1,918 | 1,512 | -21% | 0 | 0 | — |
case-13 | pass→pass | 15,494 | 2,639 | -83% | 1 | 1 | 0% | 2,576 | 1,055 | -59% | 0 | 0 | — |
case-14 | fail→pass | 16,573 | 7,963 | -52% | 1 | 1 | 0% | 2,029 | 1,241 | -39% | 0 | 0 | — |
case-15 | fail→pass | 13,909 | 7,056 | -49% | 1 | 1 | 0% | 2,084 | 1,096 | -47% | 0 | 0 | — |
case-16 | pass→pass | 4,324 | 3,884 | -10% | 1 | 1 | 0% | 642 | 1,051 | +64% | 0 | 0 | — |
case-17 | pass→pass | 11,840 | 7,026 | -41% | 1 | 1 | 0% | 1,486 | 1,829 | +23% | 0 | 0 | — |
case-18 | fail→pass | 21,813 | 4,021 | -82% | 1 | 1 | 0% | 1,716 | 1,404 | -18% | 0 | 0 | — |
case-19 | pass→pass | 26,116 | 18,974 | -27% | 1 | 1 | 0% | 3,222 | 3,555 | +10% | 0 | 0 | — |
case-20 | pass→pass | 13,308 | 10,850 | -18% | 1 | 1 | 0% | 1,837 | 2,585 | +41% | 0 | 0 | — |
case-21 | pass→pass | 16,668 | 11,720 | -30% | 1 | 1 | 0% | 2,161 | 2,786 | +29% | 0 | 0 | — |
case-22 | fail→pass | 15,489 | 12,283 | -21% | 1 | 1 | 0% | 954 | 2,072 | +117% | 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. 22 cases were attempted. The headline lift of +32 percentage points is the difference between those two pass rates over the 22 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.