Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Operate documentation-oriented Spec Kitty missions and ensure docs stay tied to shipped behavior and doctrine.
.claude/skills/priivacy-ai-spk-mission-documentation/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✓→✗ | ▼ Worse | -60% | 0% |
| case-02 | ✓→✓ | ＝ Same ✓ | -8% | 0% |
| case-03 | ✓→✓ | ＝ Same ✓ | -1% | 0% |
| case-07 | ✓→✓ | ＝ Same ✓ | -39% | 0% |
| case-08 | ✗→✗ | ＝ Same ✗ | 2% | 0% |
Use this skill when the mission is primarily documentation, release notes, guides, or user-facing explanation.
docs as acceptance criteria.
clearly marked.
spk-doctrine-glossary.spk-start-command-map.Documentation should make the next user action obvious and should not create a second source of truth for runtime behavior.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | pass→fail | 7,389 | 2,374 | -68% | 1 | 1 | 0% | 1,439 | 572 | -60% | 0 | 0 | — |
case-08 | fail→fail | 25,805 | 25,323 | -2% | 1 | 1 | 0% | 6,186 | 6,329 | +2% | 0 | 0 | — |
case-14 | fail→fail | 8,748 | 6,637 | -24% | 1 | 1 | 0% | 1,464 | 1,328 | -9% | 0 | 0 | — |
case-02 | pass→pass | 8,552 | 6,963 | -19% | 1 | 1 | 0% | 1,790 | 1,644 | -8% | 0 | 0 | — |
case-03 | pass→pass | 3,972 | 2,727 | -31% | 1 | 1 | 0% | 659 | 650 | -1% | 0 | 0 | — |
case-04 | fail→fail | 8,889 | 7,459 | -16% | 1 | 1 | 0% | 1,640 | 1,340 | -18% | 0 | 0 | — |
case-05 | fail→fail | 11,459 | 7,750 | -32% | 1 | 1 | 0% | 2,173 | 1,572 | -28% | 0 | 0 | — |
case-06 | fail→fail | 13,115 | 8,987 | -31% | 1 | 1 | 0% | 2,459 | 1,780 | -28% | 0 | 0 | — |
case-07 | pass→pass | 13,681 | 7,576 | -45% | 1 | 1 | 0% | 2,621 | 1,609 | -39% | 0 | 0 | — |
case-09 | fail→fail | 9,519 | 7,492 | -21% | 1 | 1 | 0% | 1,889 | 1,516 | -20% | 0 | 0 | — |
case-10 | fail→fail | 9,814 | 7,719 | -21% | 1 | 1 | 0% | 1,889 | 1,559 | -17% | 0 | 0 | — |
case-11 | fail→fail | 13,636 | 9,672 | -29% | 1 | 1 | 0% | 2,285 | 1,748 | -24% | 0 | 0 | — |
case-12 | fail→fail | 12,375 | 11,142 | -10% | 1 | 1 | 0% | 2,244 | 1,922 | -14% | 0 | 0 | — |
case-13 | fail→fail | 9,096 | 9,040 | -1% | 1 | 1 | 0% | 2,138 | 1,973 | -8% | 0 | 0 | — |
case-15 | fail→fail | 16,139 | 11,565 | -28% | 1 | 1 | 0% | 2,987 | 2,261 | -24% | 0 | 0 | — |
case-16 | fail→fail | 14,319 | 7,768 | -46% | 1 | 1 | 0% | 2,864 | 1,605 | -44% | 0 | 0 | — |
case-17 | fail→fail | 13,449 | 7,471 | -44% | 1 | 1 | 0% | 2,586 | 1,463 | -43% | 0 | 0 | — |
case-18 | fail→fail | 15,681 | 12,206 | -22% | 1 | 1 | 0% | 3,073 | 2,496 | -19% | 0 | 0 | — |
case-19 | fail→fail | 14,549 | 9,553 | -34% | 1 | 1 | 0% | 2,513 | 1,699 | -32% | 0 | 0 | — |
case-20 | fail→fail | 13,167 | 8,286 | -37% | 1 | 1 | 0% | 2,400 | 1,681 | -30% | 0 | 0 | — |
case-21 | fail→fail | 12,870 | 11,817 | -8% | 1 | 1 | 0% | 2,445 | 2,412 | -1% | 0 | 0 | — |
case-22 | fail→fail | 13,716 | 14,824 | +8% | 1 | 1 | 0% | 2,717 | 2,812 | +3% | 0 | 0 | — |
case-23 | fail→fail | 25,660 | 10,976 | -57% | 1 | 1 | 0% | 6,180 | 2,325 | -62% | 0 | 0 | — |
case-24 | fail→fail | 8,921 | 5,550 | -38% | 1 | 1 | 0% | 1,682 | 1,189 | -29% | 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 -4 percentage points is the difference between those two pass rates over the 24 comparable cases. 1 case got worse with the skill loaded, and it is included in that figure.
Other measured skills in the registry, with their headline benchmark lift.