obsidian
Composable Obsidian vault operations via CLI. Use when ANY task needs to read, write, search, tag, or query the Obsidian vault. This is the ONLY correct way to interact with the vault. Provides typed wrappers around the Obsidian CLI (v1.12+). Requires Obsidian to be running. Use when user says "save to vault", "obsidian", "research note", "daily note", "tag", "summarize and save", or any vault operation.
90
Quality
88%
Does it follow best practices?
Impact
93%
4.89xAverage score across 3 eval scenarios
Evaluation results
80%
45%Vault Import Script for Research Reports
PowerShell heredoc syntax and vault CLI invocation
Single-quoted heredoc used
0%
0%
No double-quoted heredoc
100%
100%
Correct script path
0%
100%
No temp Python files
100%
100%
No disk writes for vault content
100%
100%
Single pipe per note
0%
100%
Verify after create
0%
100%
Uses create command
0%
100%
Without context: $0.4133 · 1m 51s · 23 turns · 27 in / 5,942 out tokens
With context: $0.6914 · 2m 21s · 22 turns · 26 in / 7,554 out tokens
100%
100%Documenting a Production Deployment Discovery
Agent memory format and deduplication
Deduplication search first
0%
100%
Uses obsidian import
0%
100%
agent-memory tag
0%
100%
Memory category tag
0%
100%
source-skill field
0%
100%
captured field
0%
100%
confidence field
0%
100%
Context section
0%
100%
Insight section
0%
100%
Evidence section
0%
100%
Agent Memories path
0%
100%
Without context: $0.5613 · 2m 8s · 35 turns · 496 in / 6,734 out tokens
With context: $0.4206 · 1m 17s · 16 turns · 53 in / 3,971 out tokens
100%
78%Sync the Obsidian Wrapper with a New CLI Version
Obsidian CLI upgrade workflow
Run --diff first
0%
100%
Run --guide
0%
100%
Update obsidian.py
100%
100%
Update SKILL.md Quick Reference
0%
100%
Update README.md
50%
100%
Run --save after updates
0%
100%
Correct commit message format
0%
100%
Correct step ordering
40%
100%
Without context: $0.3328 · 1m 53s · 14 turns · 17 in / 6,688 out tokens
With context: $0.7545 · 2m 24s · 20 turns · 774 in / 8,148 out tokens
- Evaluated
- Agent
- Claude Code
- Model
- Claude Sonnet 4.6
Table of Contents
Is this your skill?
If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.