Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Handle Spec Kitty team authentication, hosted credentials, account selection, and auth-related recovery.
.claude/skills/priivacy-ai-spk-team-auth/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -58% | 0% |
| case-02 | ✗→✓ | ▲ Improved | -17% | 0% |
| case-04 | ✗→✓ | ▲ Improved | -53% | 0% |
| case-06 | ✗→✓ | ▲ Improved | -55% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -68% | 0% |
Use this skill when the user asks about login, hosted auth, team account access, or auth failures in sync/tracker workflows.
spk-team-tracker.spk-team-sync.inferred.
When testing hosted auth with tracker flows from the CLI on this computer, opt into hosted mode with SPEC_KITTY_ENABLE_SAAS_SYNC=1. The flag is the tracker-hosted rollout gate; it does not restore the removed standalone sync transport. Prefer writing it once into .kittify/.kitty.env (repo-scoped) or ${SPEC_KITTY_HOME}/.kitty.env (machine-wide) over a per-shell export — a shell export arms every project that shell subsequently touches, not just the one you're testing. Run spec-kitty doctor env-file to confirm which tier is active.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 17,419 | 8,680 | -50% | 1 | 1 | 0% | 2,227 | 939 | -58% | 0 | 0 | — |
case-02 | fail→pass | 18,798 | 13,275 | -29% | 1 | 1 | 0% | 2,161 | 1,789 | -17% | 0 | 0 | — |
case-03 | fail→fail | 18,623 | 12,929 | -31% | 1 | 1 | 0% | 2,147 | 1,649 | -23% | 0 | 0 | — |
case-04 | fail→pass | 13,758 | 7,816 | -43% | 1 | 1 | 0% | 1,554 | 728 | -53% | 0 | 0 | — |
case-05 | fail→fail | 12,675 | 7,883 | -38% | 1 | 1 | 0% | 1,328 | 853 | -36% | 0 | 0 | — |
case-11 | pass→pass | 12,787 | 8,789 | -31% | 1 | 1 | 0% | 1,224 | 932 | -24% | 0 | 0 | — |
case-06 | fail→pass | 20,238 | 9,422 | -53% | 1 | 1 | 0% | 2,481 | 1,105 | -55% | 0 | 0 | — |
case-07 | fail→pass | 13,763 | 6,869 | -50% | 1 | 1 | 0% | 1,610 | 515 | -68% | 0 | 0 | — |
case-08 | fail→pass | 15,377 | 7,247 | -53% | 1 | 1 | 0% | 1,667 | 654 | -61% | 0 | 0 | — |
case-09 | fail→pass | 19,542 | 7,605 | -61% | 1 | 1 | 0% | 2,091 | 688 | -67% | 0 | 0 | — |
case-10 | fail→pass | 18,567 | 11,501 | -38% | 1 | 1 | 0% | 2,219 | 1,469 | -34% | 0 | 0 | — |
case-12 | fail→pass | 11,105 | 7,219 | -35% | 1 | 1 | 0% | 999 | 661 | -34% | 0 | 0 | — |
case-13 | pass→pass | 16,250 | 9,837 | -39% | 1 | 1 | 0% | 2,069 | 1,131 | -45% | 0 | 0 | — |
case-14 | pass→pass | 18,112 | 12,491 | -31% | 1 | 1 | 0% | 2,058 | 1,487 | -28% | 0 | 0 | — |
case-15 | pass→pass | 19,334 | 12,160 | -37% | 1 | 1 | 0% | 2,376 | 1,404 | -41% | 0 | 0 | — |
case-16 | pass→pass | 15,426 | 9,655 | -37% | 1 | 1 | 0% | 1,819 | 1,153 | -37% | 0 | 0 | — |
case-17 | pass→pass | 14,350 | 7,976 | -44% | 1 | 1 | 0% | 1,434 | 767 | -47% | 0 | 0 | — |
case-18 | fail→pass | 16,866 | 7,865 | -53% | 1 | 1 | 0% | 2,320 | 743 | -68% | 0 | 0 | — |
case-19 | fail→pass | 21,579 | 10,863 | -50% | 1 | 1 | 0% | 2,818 | 1,297 | -54% | 0 | 0 | — |
case-20 | fail→pass | 12,867 | 7,237 | -44% | 1 | 1 | 0% | 1,639 | 653 | -60% | 0 | 0 | — |
case-21 | pass→pass | 15,923 | 8,825 | -45% | 1 | 1 | 0% | 1,849 | 833 | -55% | 0 | 0 | — |
case-22 | fail→pass | 15,622 | 10,962 | -30% | 1 | 1 | 0% | 1,813 | 1,323 | -27% | 0 | 0 | — |
case-23 | pass→pass | 17,726 | 19,382 | +9% | 1 | 1 | 0% | 2,364 | 3,017 | +28% | 0 | 0 | — |
case-24 | pass→pass | 14,988 | 15,068 | +1% | 1 | 1 | 0% | 1,915 | 2,237 | +17% | 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. 24 cases were attempted. The headline lift of +54 percentage points is the difference between those two pass rates over the 24 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.
| Model | Method | Date | Lift |
|---|---|---|---|
| gemini-3.6-flash | verified | 8/27/2026 | +64% |
| gemini-3.6-flash | verified | 8/13/2026 | +73% |
Other measured skills in the registry, with their headline benchmark lift.