Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Citation plugin for Obsidian note-taking with BibTeX support
.claude/skills/brycewang-stanford-obsidian-citation-guide/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-18 | ✗→✓ | ▲ Improved | 27% | 0% |
| case-02 | ✓→✓ | = Same ✓ | 19% | 0% |
| case-03 | ✓→✓ | = Same ✓ | 72% | 0% |
| case-04 | ✓→✓ | = Same ✓ | 109% | 0% |
| case-05 | ✓→✓ | = Same ✓ | 48% | 0% |
A skill for using the Obsidian Citation Plugin to integrate bibliographic references from BibTeX or CSL-JSON files into the Obsidian knowledge management workflow. Based on obsidian-citation-plugin (1K stars), this skill enables researchers to create literature notes, insert citations, and build interconnected knowledge bases grounded in their reference libraries.
Academic researchers need their notes to be deeply connected to their references. The Obsidian Citation Plugin bridges BibTeX or CSL-JSON bibliography files with Obsidian's note-taking system, enabling seamless creation of literature notes from reference entries, quick citation insertion while writing, and automatic backlinking between notes and their source references.
Unlike Zotero-specific integrations, this plugin works with any tool that exports standard BibTeX or CSL-JSON, making it compatible with Zotero, JabRef, Mendeley, EndNote, and even hand-crafted .bib files.
Prerequisites
Installation
Configuration
references/ or literature/)Template Variables
{{citekey}} - the unique citation key for the reference{{title}} - the full title of the work{{authorString}} - formatted author list{{year}} - publication year{{abstract}} - the abstract if available in the bibliography{{DOI}} - Digital Object Identifier{{URL}} - web link to the publication{{containerTitle}} - journal or conference name{{page}} - page numbers{{tags}} - keywords or tags from the bibliography entryCreating Literature Notes
Inserting Citations
[@citekey], [[@citekey]])Browsing References
Minimal Notes (Quick Reference)
Standard Notes (Working Knowledge)
Deep Notes (Core References)
Linking Strategies
Dataview Integration
Graph View Usage
Zotero to Obsidian Pipeline
Writing Pipeline
This skill enhances the Research-Claw knowledge management workflow:
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 26,042 | 50,152 | +93% | 1 | 1 | 0% | 3,738 | 6,249 | +67% | 0 | 0 | — |
case-02 | pass→pass | 32,523 | 24,497 | -25% | 1 | 1 | 0% | 5,099 | 6,053 | +19% | 0 | 0 | — |
case-03 | pass→pass | 15,283 | 12,684 | -17% | 1 | 1 | 0% | 2,047 | 3,523 | +72% | 0 | 0 | — |
case-04 | pass→pass | 7,128 | 7,180 | +1% | 1 | 1 | 0% | 1,350 | 2,815 | +109% | 0 | 0 | — |
case-05 | pass→pass | 13,156 | 13,845 | +5% | 1 | 1 | 0% | 2,438 | 3,620 | +48% | 0 | 0 | — |
case-06 | pass→pass | 12,708 | 5,218 | -59% | 1 | 1 | 0% | 1,888 | 2,427 | +29% | 0 | 0 | — |
case-07 | pass→pass | 9,223 | 8,046 | -13% | 1 | 1 | 0% | 1,600 | 2,574 | +61% | 0 | 0 | — |
case-08 | pass→pass | 10,376 | 7,685 | -26% | 1 | 1 | 0% | 1,920 | 2,564 | +34% | 0 | 0 | — |
case-09 | fail→fail | 12,452 | 9,019 | -28% | 1 | 1 | 0% | 1,688 | 2,725 | +61% | 0 | 0 | — |
case-10 | pass→pass | 13,279 | 11,746 | -12% | 1 | 1 | 0% | 2,150 | 3,056 | +42% | 0 | 0 | — |
case-11 | fail→fail | 18,559 | 17,302 | -7% | 1 | 1 | 0% | 2,860 | 4,230 | +48% | 0 | 0 | — |
case-12 | pass→pass | 14,289 | 10,677 | -25% | 1 | 1 | 0% | 1,975 | 3,263 | +65% | 0 | 0 | — |
case-13 | pass→pass | 13,394 | 10,447 | -22% | 1 | 1 | 0% | 2,334 | 2,992 | +28% | 0 | 0 | — |
case-14 | pass→pass | 17,240 | 16,448 | -5% | 1 | 1 | 0% | 3,088 | 4,425 | +43% | 0 | 0 | — |
case-15 | pass→pass | 15,710 | 16,631 | +6% | 1 | 1 | 0% | 2,476 | 3,806 | +54% | 0 | 0 | — |
case-16 | pass→pass | 19,624 | 10,456 | -47% | 1 | 1 | 0% | 2,341 | 3,331 | +42% | 0 | 0 | — |
case-17 | pass→pass | 14,327 | 8,554 | -40% | 1 | 1 | 0% | 1,957 | 2,964 | +51% | 0 | 0 | — |
case-18 | fail→pass | 22,121 | 13,017 | -41% | 1 | 1 | 0% | 2,694 | 3,418 | +27% | 0 | 0 | — |
case-19 | pass→pass | 20,231 | 20,496 | +1% | 1 | 1 | 0% | 2,890 | 4,387 | +52% | 0 | 0 | — |
case-20 | pass→pass | 14,214 | 20,065 | +41% | 1 | 1 | 0% | 2,328 | 4,658 | +100% | 0 | 0 | — |
case-21 | pass→pass | 18,252 | 14,124 | -23% | 1 | 1 | 0% | 2,675 | 3,909 | +46% | 0 | 0 | — |
case-22 | pass→pass | 13,095 | 12,037 | -8% | 1 | 1 | 0% | 2,057 | 3,419 | +66% | 0 | 0 | — |
case-23 | pass→pass | 10,346 | 5,685 | -45% | 1 | 1 | 0% | 1,350 | 2,201 | +63% | 0 | 0 | — |
case-24 | pass→pass | 12,152 | 5,113 | -58% | 1 | 1 | 0% | 1,775 | 2,258 | +27% | 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.
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.