Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Tactic: Full attack lifecycle — threat surface enumeration, attack vector generation, systematic probing, and finding aggregation across all surfaces.
.claude/skills/yogsoth-ai-structured-attack-campaign/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-07 | ✗→✓ | ▲ Improved | -14% | 0% |
| case-04 | ✗→✓ | ▲ Improved | 373% | 0% |
| case-10 | ✗→✓ | ▲ Improved | -33% | 0% |
| case-13 | ✗→✓ | ▲ Improved | -41% | 0% |
| case-14 | ✗→✓ | ▲ Improved | -1% | 0% |
Complete attack lifecycle from surface enumeration through probing to aggregated findings.
<!-- BEGIN available-tables (generated) -->
Optional, no fixed order; the final leaf is always a sop.
| SOP | When to use | | --- | --- | | attack-resilience-scoring | Compute overall resilience score (0.0-1.0) based on attack results, coverage, and vulnerability severity distribution. | | attack-vector-generation | Generate specific attack strategies for a given threat surface, producing concrete probes that can be executed. | | finding-aggregation | Aggregate, deduplicate, and classify findings from multiple probes into a coherent vulnerability report. | | probe-execution | Execute a single attack probe against an artifact, record the result with evidence and severity classification. | | threat-surface-mapping | Enumerate all attackable surfaces of an artifact — logical, empirical, methodological, social, and practical dimensions. |
<!-- END available-tables (generated) -->
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-07 | fail→pass | 12,619 | 6,477 | -49% | 1 | 1 | 0% | 1,200 | 1,031 | -14% | 0 | 0 | — |
case-06 | pass→pass | 24,212 | 26,001 | +7% | 1 | 1 | 0% | 3,174 | 3,785 | +19% | 0 | 0 | — |
case-01 | fail→fail | 16,446 | 16,368 | -0% | 1 | 1 | 0% | 1,505 | 1,557 | +3% | 0 | 0 | — |
case-02 | fail→fail | 8,623 | 13,295 | +54% | 1 | 1 | 0% | 818 | 1,323 | +62% | 0 | 0 | — |
case-03 | fail→fail | 10,351 | 16,228 | +57% | 1 | 1 | 0% | 1,203 | 1,821 | +51% | 0 | 0 | — |
case-04 | fail→pass | 6,727 | 48,314 | +618% | 1 | 1 | 0% | 1,453 | 6,869 | +373% | 0 | 0 | — |
case-05 | fail→fail | 7,773 | 19,220 | +147% | 1 | 1 | 0% | 266 | 1,630 | +513% | 0 | 0 | — |
case-08 | pass→pass | 7,386 | 8,143 | +10% | 1 | 1 | 0% | 1,108 | 974 | -12% | 0 | 0 | — |
case-09 | pass→pass | 19,665 | 3,665 | -81% | 1 | 1 | 0% | 2,033 | 1,039 | -49% | 0 | 0 | — |
case-10 | fail→pass | 15,165 | 8,628 | -43% | 1 | 1 | 0% | 1,515 | 1,017 | -33% | 0 | 0 | — |
case-11 | pass→pass | 21,507 | 9,037 | -58% | 1 | 1 | 0% | 2,545 | 1,808 | -29% | 0 | 0 | — |
case-12 | pass→pass | 20,630 | 22,870 | +11% | 1 | 1 | 0% | 2,735 | 2,297 | -16% | 0 | 0 | — |
case-13 | fail→pass | 11,471 | 15,740 | +37% | 1 | 1 | 0% | 1,770 | 1,041 | -41% | 0 | 0 | — |
case-14 | fail→pass | 23,046 | 13,629 | -41% | 1 | 1 | 0% | 1,431 | 1,411 | -1% | 0 | 0 | — |
case-15 | fail→pass | 20,658 | 7,171 | -65% | 1 | 1 | 0% | 2,217 | 729 | -67% | 0 | 0 | — |
case-16 | fail→pass | 25,625 | 15,108 | -41% | 1 | 1 | 0% | 2,979 | 1,897 | -36% | 0 | 0 | — |
case-17 | fail→pass | 19,311 | 13,292 | -31% | 1 | 1 | 0% | 2,262 | 1,775 | -22% | 0 | 0 | — |
case-18 | pass→pass | 17,735 | 13,985 | -21% | 1 | 1 | 0% | 3,036 | 2,717 | -11% | 0 | 0 | — |
case-19 | fail→pass | 13,995 | 7,195 | -49% | 1 | 1 | 0% | 1,335 | 810 | -39% | 0 | 0 | — |
case-20 | pass→pass | 20,151 | 4,752 | -76% | 1 | 1 | 0% | 2,321 | 1,342 | -42% | 0 | 0 | — |
case-21 | pass→pass | 21,335 | 17,318 | -19% | 1 | 1 | 0% | 2,418 | 2,494 | +3% | 0 | 0 | — |
case-22 | fail→pass | 27,582 | 7,060 | -74% | 1 | 1 | 0% | 1,229 | 708 | -42% | 0 | 0 | — |
case-23 | fail→pass | 19,521 | 7,681 | -61% | 1 | 1 | 0% | 1,832 | 874 | -52% | 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, and 21 counted toward the lift figure. The other 2 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +48 percentage points is the difference between those two pass rates over the 21 comparable cases. 1 case got worse with the skill loaded, and it is included in that figure.
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.