Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Validates land titles against the Cameroon digital registry to prevent fraud and verify ownership.
.claude/skills/thomasmoreai-land-title-verifier/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-09 | ✗→✓ | ▲ Improved | -22% | 0% |
| case-20 | ✓→✓ | ＝ Same ✓ | -21% | 0% |
| case-21 | ✓→✓ | ＝ Same ✓ | -26% | 0% |
| case-22 | ✓→✓ | ＝ Same ✓ | -21% | 0% |
| case-01 | ✗→✗ | ＝ Same ✗ | -56% | 0% |
This skill allows the agent to check the authenticity of a "Titre Foncier" (Land Title) in Cameroon. It should be used whenever a user provides a title number or asks for property verification.
check_title: Queries the digital land registry database.title_number: (string) The unique ID of the land title (e.g., "1234/LIT").region: (string) Optional. The administrative region, such as "Littoral", "Centre", or "Ouest".get_title_history: Return the chain of previous owners for a given title number.title_number: (invalid_type) The unique ID of the land title (e.g., "1234/LIT").Superficie" (Area in square meters).
suffix), flag it as "Suspicious" and suggest a physical "Concierge Verification."
Certificat de Propriété") but keep the summary in the user's preferred language.
User: "Can you check land title 5678/LIT in Douala?" Action: Call check_title(title_number="5678/LIT", region="Littoral")
User: "I want to see the history of this property: 9912/CEN." Action: Call get_title_history(title_number="9912/CEN")
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 9,750 | 4,987 | -49% | 1 | 1 | 0% | 1,720 | 759 | -56% | 0 | 0 | — |
case-02 | fail→fail | 5,600 | 6,916 | +24% | 1 | 1 | 0% | 964 | 874 | -9% | 0 | 0 | — |
case-12 | fail→fail | 2,850 | 6,547 | +130% | 1 | 1 | 0% | 459 | 1,044 | +127% | 0 | 0 | — |
case-03 | fail→fail | 8,491 | 8,017 | -6% | 1 | 1 | 0% | 1,570 | 1,159 | -26% | 0 | 0 | — |
case-04 | fail→fail | 6,240 | 4,280 | -31% | 1 | 1 | 0% | 1,114 | 681 | -39% | 0 | 0 | — |
case-05 | fail→fail | 10,338 | 4,527 | -56% | 1 | 1 | 0% | 1,962 | 733 | -63% | 0 | 0 | — |
case-06 | fail→fail | 11,137 | 7,632 | -31% | 1 | 1 | 0% | 2,156 | 963 | -55% | 0 | 0 | — |
case-07 | fail→fail | 7,716 | 4,706 | -39% | 1 | 1 | 0% | 1,416 | 791 | -44% | 0 | 0 | — |
case-08 | fail→fail | 9,909 | 4,614 | -53% | 1 | 1 | 0% | 1,946 | 701 | -64% | 0 | 0 | — |
case-09 | fail→pass | 11,604 | 13,117 | +13% | 1 | 1 | 0% | 1,949 | 1,515 | -22% | 0 | 0 | — |
case-10 | fail→fail | 7,659 | 3,907 | -49% | 1 | 1 | 0% | 1,404 | 733 | -48% | 0 | 0 | — |
case-11 | fail→fail | 4,335 | 4,381 | +1% | 1 | 1 | 0% | 685 | 681 | -1% | 0 | 0 | — |
case-13 | fail→fail | 9,554 | 4,047 | -58% | 1 | 1 | 0% | 1,693 | 602 | -64% | 0 | 0 | — |
case-14 | fail→fail | 4,344 | 4,061 | -7% | 1 | 1 | 0% | 721 | 687 | -5% | 0 | 0 | — |
case-15 | fail→fail | 4,070 | 4,711 | +16% | 1 | 1 | 0% | 675 | 619 | -8% | 0 | 0 | — |
case-16 | fail→fail | 10,494 | 10,657 | +2% | 1 | 1 | 0% | 2,010 | 1,452 | -28% | 0 | 0 | — |
case-17 | fail→fail | 7,719 | 4,707 | -39% | 1 | 1 | 0% | 1,195 | 679 | -43% | 0 | 0 | — |
case-18 | fail→fail | 4,601 | 4,352 | -5% | 1 | 1 | 0% | 814 | 706 | -13% | 0 | 0 | — |
case-19 | fail→fail | 7,784 | 1,877 | -76% | 1 | 1 | 0% | 1,396 | 713 | -49% | 0 | 0 | — |
case-20 | pass→pass | 13,125 | 7,681 | -41% | 1 | 1 | 0% | 2,462 | 1,942 | -21% | 0 | 0 | — |
case-21 | pass→pass | 13,923 | 7,837 | -44% | 1 | 1 | 0% | 2,557 | 1,884 | -26% | 0 | 0 | — |
case-22 | pass→pass | 14,250 | 9,259 | -35% | 1 | 1 | 0% | 2,641 | 2,075 | -21% | 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 5 counted toward the lift figure. The other 17 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 +5 percentage points is the difference between those two pass rates over the 5 comparable cases. 3 cases got worse with the skill loaded, and they are 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.