Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Show ranked attack surface for a target. Invokes recon-ranker agent. Usage: /surface target.com
.claude/skills/h-mmer-surface/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-07 | ✗→✓ | ▲ Improved | 13% | 0% |
| case-08 | ✗→✓ | ▲ Improved | -10% | 0% |
| case-10 | ✗→✓ | ▲ Improved | -10% | 0% |
| case-14 | ✗→✓ | ▲ Improved | -42% | 0% |
| case-15 | ✗→✓ | ▲ Improved | -51% | 0% |
Rank attack surface for: $ARGUMENTS
recon-ranker agent: "Rank the attack surface for $ARGUMENTS. Read recon/ for discovery data and brain for tested endpoints. Output P1/P2/Kill ranking."/hunt $ARGUMENTS to start testing P1 targets.Rank by exploit economics.
P1 requires at least two of:
Kill or P3 assets that are static marketing pages, hardened vendor panels with no program-owned data, or endpoints already exhausted with strong evidence. Every P1 must include the best first vuln class and first request to try.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 14,545 | 14,946 | +3% | 1 | 1 | 0% | 2,552 | 1,735 | -32% | 0 | 0 | — |
case-02 | fail→fail | 15,808 | 4,902 | -69% | 1 | 1 | 0% | 2,803 | 605 | -78% | 0 | 0 | — |
case-03 | fail→fail | 7,935 | 12,580 | +59% | 1 | 1 | 0% | 1,089 | 1,757 | +61% | 0 | 0 | — |
case-04 | fail→fail | 11,416 | 11,278 | -1% | 1 | 1 | 0% | 1,217 | 1,398 | +15% | 0 | 0 | — |
case-05 | fail→fail | 18,312 | 12,790 | -30% | 1 | 1 | 0% | 1,944 | 1,845 | -5% | 0 | 0 | — |
case-06 | pass→pass | 10,645 | 9,641 | -9% | 1 | 1 | 0% | 1,773 | 2,086 | +18% | 0 | 0 | — |
case-07 | fail→pass | 11,038 | 11,716 | +6% | 1 | 1 | 0% | 1,834 | 2,070 | +13% | 0 | 0 | — |
case-08 | fail→pass | 11,803 | 8,234 | -30% | 1 | 1 | 0% | 1,911 | 1,728 | -10% | 0 | 0 | — |
case-09 | pass→pass | 10,601 | 8,436 | -20% | 1 | 1 | 0% | 1,680 | 1,163 | -31% | 0 | 0 | — |
case-10 | fail→pass | 7,595 | 5,063 | -33% | 1 | 1 | 0% | 1,186 | 1,071 | -10% | 0 | 0 | — |
case-11 | pass→pass | 16,049 | 15,692 | -2% | 1 | 1 | 0% | 1,398 | 1,652 | +18% | 0 | 0 | — |
case-12 | fail→fail | 11,019 | 11,644 | +6% | 1 | 1 | 0% | 1,445 | 1,608 | +11% | 0 | 0 | — |
case-13 | fail→fail | 9,016 | 2,340 | -74% | 1 | 1 | 0% | 1,420 | 636 | -55% | 0 | 0 | — |
case-14 | fail→pass | 9,814 | 3,593 | -63% | 1 | 1 | 0% | 1,678 | 971 | -42% | 0 | 0 | — |
case-15 | fail→pass | 11,670 | 3,854 | -67% | 1 | 1 | 0% | 1,977 | 976 | -51% | 0 | 0 | — |
case-16 | fail→pass | 9,880 | 4,118 | -58% | 1 | 1 | 0% | 1,579 | 1,041 | -34% | 0 | 0 | — |
case-17 | pass→pass | 9,259 | 6,628 | -28% | 1 | 1 | 0% | 1,509 | 1,540 | +2% | 0 | 0 | — |
case-18 | fail→pass | 9,307 | 1,522 | -84% | 1 | 1 | 0% | 1,534 | 507 | -67% | 0 | 0 | — |
case-19 | pass→pass | 9,852 | 6,053 | -39% | 1 | 1 | 0% | 1,648 | 1,379 | -16% | 0 | 0 | — |
case-20 | pass→pass | 5,491 | 4,017 | -27% | 1 | 1 | 0% | 1,120 | 1,006 | -10% | 0 | 0 | — |
case-21 | pass→pass | 13,690 | 11,638 | -15% | 1 | 1 | 0% | 2,483 | 2,523 | +2% | 0 | 0 | — |
case-22 | pass→pass | 17,224 | 15,575 | -10% | 1 | 1 | 0% | 3,279 | 2,284 | -30% | 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, and 21 counted toward the lift figure. The other 1 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 +32 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.