Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Sync program scope, policy, and hacktivity from a bug bounty platform. Usage: /sync hackerone tesla or /sync bugcrowd uber
.claude/skills/h-mmer-sync/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-04 | ✗→✓ | ▲ Improved | -38% | 0% |
| case-11 | ✗→✓ | ▲ Improved | 2% | 0% |
| case-15 | ✗→✓ | ▲ Improved | -26% | 0% |
| case-16 | ✗→✓ | ▲ Improved | 14% | 0% |
| case-17 | ✗→✓ | ▲ Improved | 17% | 0% |
Sync bug bounty program data: $ARGUMENTS
Parse the arguments as: <platform> <program_handle>
bounty-platforms MCP server tool sync_program with the platform and program handle. This fetches scope, policy, and hacktivity and writes them to the current directory.uv run python3 ../../tools/brain.py init if brain isn't initialized yet.scope.yaml and hacktivity.md files.uv run python3 ../../tools/brain.py log "Synced program data from <platform>/<program>"Policy is hunting input, not paperwork.
Extract and persist:
End with a hunt bias: where the program appears to pay, where it appears saturated, and what proof standard the policy implies.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-09 | fail→fail | 5,678 | 5,825 | +3% | 1 | 1 | 0% | 797 | 699 | -12% | 0 | 0 | — |
case-01 | fail→fail | 14,079 | 6,483 | -54% | 1 | 1 | 0% | 2,295 | 739 | -68% | 0 | 0 | — |
case-02 | fail→fail | 14,385 | 5,309 | -63% | 1 | 1 | 0% | 2,270 | 671 | -70% | 0 | 0 | — |
case-03 | fail→fail | 12,619 | 4,889 | -61% | 1 | 1 | 0% | 1,962 | 523 | -73% | 0 | 0 | — |
case-04 | fail→pass | 17,358 | 16,043 | -8% | 1 | 1 | 0% | 2,580 | 1,593 | -38% | 0 | 0 | — |
case-05 | pass→pass | 8,122 | 32,057 | +295% | 1 | 1 | 0% | 1,487 | 2,370 | +59% | 0 | 0 | — |
case-06 | fail→fail | 7,626 | 2,007 | -74% | 1 | 1 | 0% | 1,148 | 674 | -41% | 0 | 0 | — |
case-07 | fail→fail | 9,295 | 5,484 | -41% | 1 | 1 | 0% | 1,866 | 662 | -65% | 0 | 0 | — |
case-08 | fail→fail | 14,038 | 10,105 | -28% | 1 | 1 | 0% | 2,465 | 1,203 | -51% | 0 | 0 | — |
case-10 | fail→fail | 13,415 | 6,451 | -52% | 1 | 1 | 0% | 2,359 | 722 | -69% | 0 | 0 | — |
case-11 | fail→pass | 10,742 | 8,554 | -20% | 1 | 1 | 0% | 1,655 | 1,691 | +2% | 0 | 0 | — |
case-12 | pass→pass | 12,609 | 10,797 | -14% | 1 | 1 | 0% | 2,107 | 1,961 | -7% | 0 | 0 | — |
case-13 | pass→pass | 13,327 | 17,218 | +29% | 1 | 1 | 0% | 2,201 | 1,926 | -12% | 0 | 0 | — |
case-14 | pass→pass | 14,571 | 12,521 | -14% | 1 | 1 | 0% | 2,485 | 2,340 | -6% | 0 | 0 | — |
case-15 | fail→pass | 20,014 | 8,175 | -59% | 1 | 1 | 0% | 2,142 | 1,591 | -26% | 0 | 0 | — |
case-16 | fail→pass | 11,112 | 10,946 | -1% | 1 | 1 | 0% | 1,846 | 2,110 | +14% | 0 | 0 | — |
case-17 | fail→pass | 11,724 | 12,426 | +6% | 1 | 1 | 0% | 2,034 | 2,374 | +17% | 0 | 0 | — |
case-18 | fail→pass | 9,676 | 2,873 | -70% | 1 | 1 | 0% | 1,738 | 912 | -48% | 0 | 0 | — |
case-19 | fail→pass | 11,100 | 3,472 | -69% | 1 | 1 | 0% | 1,868 | 904 | -52% | 0 | 0 | — |
case-20 | fail→fail | 14,166 | 10,154 | -28% | 1 | 1 | 0% | 2,162 | 1,825 | -16% | 0 | 0 | — |
case-21 | fail→pass | 11,849 | 14,900 | +26% | 1 | 1 | 0% | 1,862 | 2,594 | +39% | 0 | 0 | — |
case-22 | pass→pass | 15,279 | 27,637 | +81% | 1 | 1 | 0% | 2,740 | 5,390 | +97% | 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 15 counted toward the lift figure. The other 7 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 +36 percentage points is the difference between those two pass rates over the 15 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.