Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Strategy for identifying and typing relationships between concepts — create edges with appropriate types and weights.
.claude/skills/yogsoth-ai-relation-typing/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-02 | ✗→✓ | ▲ Improved | -25% | 0% |
| case-08 | ✗→✓ | ▲ Improved | -52% | 0% |
| case-01 | ✗→✓ | ▲ Improved | -33% | 0% |
| case-04 | ✗→✓ | ▲ Improved | 334% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -47% | 0% |
Identify relationships between extracted concepts and type them using the edge type vocabulary. Every concept should connect to at least 2 others.
Relationships are more important than concepts in an ontology. Focus on making the graph dense and meaningful. Choose edge types carefully — a wrong type is worse than a missing edge.
| Metric | S | M | L | |--------|---|---|---| | Edges created | 20 | 60 | 150 | | Edge types used | 4 | 6 | 8 | | Orphans remaining | <3 | <5 | <8 |
| Metric | Target | Current | Status |
|---------------------|--------|---------|--------|
| Edges created | X | 0 | ⬜ |
| Edge types used | X | 0 | ⬜ |
| Orphans remaining | <X | ? | ⬜ |<HARD-GATE> Cannot exit until 80% of budget met. Print state ledger before each iteration decision. </HARD-GATE>
<!-- BEGIN available-tables (generated) -->
Optional, no fixed order; the final leaf is always a sop.
| Tactic | When to use | | --- | --- | | hierarchy-construction | Tactic for building is-a and part-of hierarchies — establish parent-child relationships, verify transitivity, detect cycles. | | knowledge-structuring-consistency-checking | Tactic for verifying ontology consistency — detect contradictions, cycles, orphans, and type violations. |
Optional, no fixed order; the final leaf is always a sop.
| SOP | When to use | | --- | --- | | edge-batch-creation | SOP for creating multiple edges in a batch — efficient bulk relationship creation. |
<!-- END available-tables (generated) -->
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-02 | fail→pass | 38,771 | 30,849 | -20% | 1 | 1 | 0% | 8,038 | 6,033 | -25% | 0 | 0 | — |
case-08 | fail→pass | 26,465 | 9,029 | -66% | 1 | 1 | 0% | 1,796 | 861 | -52% | 0 | 0 | — |
case-01 | fail→pass | 39,112 | 26,498 | -32% | 1 | 1 | 0% | 8,264 | 5,508 | -33% | 0 | 0 | — |
case-03 | fail→fail | 43,923 | 37,523 | -15% | 1 | 1 | 0% | 8,253 | 7,868 | -5% | 0 | 0 | — |
case-04 | fail→pass | 9,968 | 32,216 | +223% | 1 | 1 | 0% | 1,431 | 6,216 | +334% | 0 | 0 | — |
case-05 | fail→pass | 38,187 | 8,164 | -79% | 1 | 1 | 0% | 1,843 | 979 | -47% | 0 | 0 | — |
case-06 | fail→pass | 21,875 | 10,054 | -54% | 1 | 1 | 0% | 2,632 | 1,212 | -54% | 0 | 0 | — |
case-07 | fail→pass | 39,092 | 10,555 | -73% | 1 | 1 | 0% | 1,376 | 1,070 | -22% | 0 | 0 | — |
case-09 | pass→pass | 18,472 | 18,584 | +1% | 1 | 1 | 0% | 2,025 | 2,645 | +31% | 0 | 0 | — |
case-10 | fail→pass | 12,416 | 12,762 | +3% | 1 | 1 | 0% | 2,084 | 1,490 | -29% | 0 | 0 | — |
case-11 | pass→pass | 7,711 | 12,848 | +67% | 1 | 1 | 0% | 1,355 | 1,260 | -7% | 0 | 0 | — |
case-12 | fail→pass | 14,018 | 8,632 | -38% | 1 | 1 | 0% | 2,650 | 1,408 | -47% | 0 | 0 | — |
case-13 | fail→pass | 12,942 | 6,636 | -49% | 1 | 1 | 0% | 2,470 | 1,001 | -59% | 0 | 0 | — |
case-14 | pass→pass | 15,141 | 2,716 | -82% | 1 | 1 | 0% | 1,757 | 747 | -57% | 0 | 0 | — |
case-15 | pass→pass | 22,952 | 13,732 | -40% | 1 | 1 | 0% | 2,213 | 2,238 | +1% | 0 | 0 | — |
case-16 | pass→pass | 11,949 | 7,484 | -37% | 1 | 1 | 0% | 1,614 | 749 | -54% | 0 | 0 | — |
case-17 | fail→pass | 10,764 | 4,250 | -61% | 1 | 1 | 0% | 1,795 | 743 | -59% | 0 | 0 | — |
case-18 | fail→pass | 12,008 | 3,864 | -68% | 1 | 1 | 0% | 1,699 | 647 | -62% | 0 | 0 | — |
case-19 | fail→pass | 12,595 | 5,304 | -58% | 1 | 1 | 0% | 1,748 | 674 | -61% | 0 | 0 | — |
case-20 | pass→pass | 12,132 | 36,000 | +197% | 1 | 1 | 0% | 2,186 | 8,679 | +297% | 0 | 0 | — |
case-21 | fail→pass | 30,936 | 2,846 | -91% | 1 | 1 | 0% | 2,517 | 782 | -69% | 0 | 0 | — |
case-22 | pass→fail | 14,899 | 25,226 | +69% | 1 | 1 | 0% | 929 | 4,853 | +422% | 0 | 0 | — |
case-23 | pass→fail | 8,487 | 13,166 | +55% | 1 | 1 | 0% | 979 | 2,698 | +176% | 0 | 0 | — |
case-24 | pass→fail | 5,667 | 27,974 | +394% | 1 | 1 | 0% | 1,309 | 5,541 | +323% | 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, and 22 counted toward the lift figure. The other 2 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 +46 percentage points is the difference between those two pass rates over the 22 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.