Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →9 knowledge graphs skills. Trigger: building knowledge graphs, connecting concepts, ontology design. Design: graph construction, traversal, and visualization for research knowledge.
.claude/skills/brycewang-stanford-knowledge-graph-skills/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-14 | ✗→✓ | ▲ Improved | -35% | 0% |
| case-18 | ✗→✓ | ▲ Improved | -16% | 0% |
| case-19 | ✗→✓ | ▲ Improved | -35% | 0% |
| case-21 | ✗→✓ | ▲ Improved | -48% | 0% |
| case-22 | ✗→✓ | ▲ Improved | -15% | 0% |
Select the skill matching the user's need, then read its SKILL.md.
| Skill | Description | |-------|-------------| | citation-network-builder | Build and analyze citation networks from academic reference data | | concept-map-generator | Generate structured concept maps from academic texts automatically | | graphiti-guide | Build real-time knowledge graphs for AI agents using Graphiti by Zep | | knowledge-graph-construction | Build research knowledge graphs for literature synthesis and RAG systems | | notero-zotero-notion-guide | Sync Zotero references and annotations to Notion databases | | ontology-design-guide | Design ontologies and knowledge graphs for research data modeling | | openspg-guide | Ant Group knowledge graph engine with SPG and KAG framework | | rag-methodology-guide | RAG architecture for academic knowledge retrieval and synthesis | | zotero-markdb-connect-guide | Sync Zotero references to Obsidian and Logseq markdown |
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-04 | fail→fail | 43,753 | 4,479 | -90% | 1 | 1 | 0% | 2,098 | 772 | -63% | 0 | 0 | — |
case-03 | fail→fail | 45,738 | 4,706 | -90% | 1 | 1 | 0% | 2,080 | 634 | -70% | 0 | 0 | — |
case-01 | fail→fail | 49,244 | 87,098 | +77% | 1 | 1 | 0% | 3,448 | 4,385 | +27% | 0 | 0 | — |
case-02 | pass→fail | 17,046 | 45,839 | +169% | 1 | 1 | 0% | 2,336 | 738 | -68% | 0 | 0 | — |
case-05 | fail→fail | 16,754 | 33,975 | +103% | 1 | 1 | 0% | 2,639 | 655 | -75% | 0 | 0 | — |
case-06 | fail→fail | 15,345 | 6,448 | -58% | 1 | 1 | 0% | 2,378 | 784 | -67% | 0 | 0 | — |
case-07 | pass→fail | 40,555 | 5,251 | -87% | 1 | 1 | 0% | 2,388 | 871 | -64% | 0 | 0 | — |
case-08 | pass→fail | 12,024 | 36,967 | +207% | 1 | 1 | 0% | 2,089 | 736 | -65% | 0 | 0 | — |
case-09 | pass→fail | 11,337 | 34,297 | +203% | 1 | 1 | 0% | 1,882 | 731 | -61% | 0 | 0 | — |
case-10 | pass→pass | 13,098 | 7,140 | -45% | 1 | 1 | 0% | 2,244 | 1,755 | -22% | 0 | 0 | — |
case-11 | fail→fail | 14,395 | 6,285 | -56% | 1 | 1 | 0% | 2,309 | 735 | -68% | 0 | 0 | — |
case-12 | pass→pass | 13,463 | 6,318 | -53% | 1 | 1 | 0% | 2,035 | 887 | -56% | 0 | 0 | — |
case-13 | fail→fail | 10,189 | 4,481 | -56% | 1 | 1 | 0% | 1,677 | 710 | -58% | 0 | 0 | — |
case-14 | fail→pass | 8,623 | 2,723 | -68% | 1 | 1 | 0% | 1,303 | 851 | -35% | 0 | 0 | — |
case-15 | pass→fail | 24,284 | 14,966 | -38% | 1 | 1 | 0% | 4,024 | 743 | -82% | 0 | 0 | — |
case-16 | fail→fail | 18,805 | 6,684 | -64% | 1 | 1 | 0% | 2,759 | 866 | -69% | 0 | 0 | — |
case-17 | fail→fail | 12,040 | 5,059 | -58% | 1 | 1 | 0% | 2,098 | 627 | -70% | 0 | 0 | — |
case-18 | fail→pass | 7,813 | 9,478 | +21% | 1 | 1 | 0% | 1,412 | 1,190 | -16% | 0 | 0 | — |
case-19 | fail→pass | 7,986 | 3,015 | -62% | 1 | 1 | 0% | 1,297 | 843 | -35% | 0 | 0 | — |
case-20 | pass→fail | 20,620 | 6,672 | -68% | 1 | 1 | 0% | 2,452 | 585 | -76% | 0 | 0 | — |
case-21 | fail→pass | 10,648 | 4,276 | -60% | 1 | 1 | 0% | 1,921 | 1,008 | -48% | 0 | 0 | — |
case-22 | fail→pass | 9,977 | 5,681 | -43% | 1 | 1 | 0% | 1,412 | 1,202 | -15% | 0 | 0 | — |
case-23 | pass→pass | 17,407 | 11,057 | -36% | 1 | 1 | 0% | 3,538 | 2,236 | -37% | 0 | 0 | — |
case-24 | fail→pass | 11,418 | 8,283 | -27% | 1 | 1 | 0% | 1,682 | 1,509 | -10% | 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 9 counted toward the lift figure. The other 15 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 0 percentage points is the difference between those two pass rates over the 9 comparable cases. 8 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.