OBC — Obsidian Commands

30 vault commands for AI-augmented personal knowledge management. Every command reads from an Obsidian vault in real time using the obsidian CLI at /usr/local/bin/obsidian.

Requirements

• Obsidian must be running (the CLI talks to the running app)
• obsidian CLI installed at /usr/local/bin/obsidian
• Some commands integrate with Google Calendar, Google Tasks, and Gmail MCPs (optional)

How This Skill Works

1. Match the user's request to a command using the router below
2. Run shared scripts from scripts/ to gather structural data (avoid repeating the same 5-6 CLI commands in every prompt)
3. Read the prompt file from the vault at references/Prompt - <Name>.md for full command instructions
4. Execute the command following the prompt's step-by-step workflow

The prompt files in the vault are the source of truth. This skill is the orchestration layer — routing, shared tooling, workflow chains, and agents.

Command Router

Daily Operations — Routine workflows

Command	Vault File	When to Use
/context	Prompt - Context.md	Before starting any work — loads full vault context
/today	Prompt - Today.md	Morning planning — calendar + tasks + vault patterns
/close-day	Prompt - Close Day.md	Evening — extract, categorize, file insights from today
/schedule	Prompt - Schedule.md	Scheduling decisions — evaluates against priorities
/7plan	Prompt - 7plan.md	Weekly planning — reshapes 7 days around live threads

Vault Intelligence — Discovery and structural analysis

Command	Vault File	When to Use
/map	Prompt - Map.md	Topological view — clusters, themes, gaps, shape of thinking
/emerge	Prompt - Emerge.md	Surface ideas the vault implies but never states
/connect	Prompt - Connect.md	Find bridges between two specific domains
/contradict	Prompt - Contradict.md	Find incompatible beliefs held simultaneously
/drift	Prompt - Drift.md	Compare stated intentions vs. actual behavior
/backlinks	Prompt - Backlinks.md	Wire missing connections in the vault graph
/graduate	Prompt - Graduate.md	Promote daily note ideas to standalone notes
/graph	Prompt - Graph.md	Quantitative graph analytics — centrality, communities
/audit	Prompt - Audit.md	Vault health check — KPIs + prioritized action plan

Thinking Tools — Reflection, challenge, temporal analysis

Command	Vault File	When to Use
/ghost	Prompt - Ghost.md	Answer a question AS the vault's author
/trace	Prompt - Trace.md	Track how an idea evolved over weeks/months
/challenge	Prompt - Challenge.md	Pressure-test beliefs using vault evidence
/stranger	Prompt - Stranger.md	Portrait of who you are from an outsider reading cold
/compound	Prompt - Compound.md	Same question at 3 time periods — shows context compounding

Creation & Output — Producing content from vault material

Command	Vault File	When to Use
/ideas	Prompt - Ideas.md	Generate ideas across domains from vault patterns
/learned	Prompt - Learned.md	Turn a topic into writing at 3 levels (post/essay/universal)
/weekly-learnings	Prompt - Weekly Learnings.md	Compile week's insights for a team email
/make	Prompt - Make.md	Score vault ideas on readiness to become real work
/xdaily	Prompt - XDaily.md	Pull X/Twitter posts into daily notes
/xarticle	Prompt - XArticle.md	Find what to write next — graph density x energy x synthesis
/synthesize	Prompt - Synthesize.md	Weave 3+ vault threads into coherent narratives
/evolve	Prompt - Evolve.md	Find ideas at the intersection of readiness x leverage

Strategy — Big-picture decisions

Command	Vault File	When to Use
/money	Prompt - Money.md	Revenue diagnosis — asset inventory, opportunities, builds
/leverage	Prompt - Leverage.md	Find 3-7 skills where concentrated investment = breakthrough
/guests	Prompt - Guests.md	Derive podcast guests from questions the vault is asking

Shared Scripts

Every command reuses the same building blocks. Run these scripts from ${SKILL_DIR}/scripts/ to avoid repeating boilerplate.

vault-scan.sh

Structural overview used by almost every command. Outputs vault stats, top tags, orphans, deadends, unresolved links.

bash "${SKILL_DIR}/scripts/vault-scan.sh"

daily-reader.sh <N>

Read the past N days of daily notes. Outputs today's note + N-1 previous days.

bash "${SKILL_DIR}/scripts/daily-reader.sh" 7    # Past 7 days
bash "${SKILL_DIR}/scripts/daily-reader.sh" 14   # Past 14 days
bash "${SKILL_DIR}/scripts/daily-reader.sh" 30   # Past 30 days

hub-detect.sh <N>

Find the top N most-connected notes by backlink count. Outputs JSON with note name and count.

bash "${SKILL_DIR}/scripts/hub-detect.sh" 15     # Top 15 hubs

context-loader.sh

Find and read all context files. Discovers them by searching for "Context" in filenames.

bash "${SKILL_DIR}/scripts/context-loader.sh"

thread-search.sh <pattern>

Search daily notes for intention/belief/decision patterns. Used by Contradict, Challenge, Drift, Emerge, Leverage.

bash "${SKILL_DIR}/scripts/thread-search.sh" "I believe"
bash "${SKILL_DIR}/scripts/thread-search.sh" "decided|chose|going to"

temporal-track.sh <command-name>

Check for prior runs of a command in daily notes. Used by Ideas, Make, Money, Leverage, XArticle, Guests, Synthesize, Evolve, Audit.

bash "${SKILL_DIR}/scripts/temporal-track.sh" "/ideas"

Execution Pattern

For any command:

1. Run vault-scan.sh unless the command explicitly skips it (XDaily, Weekly Learnings)
2. Run daily-reader.sh N with N appropriate to the command (7 for daily ops, 14-21 for vault intelligence, 30+ for strategy)
3. Read the prompt file from the vault: Read file="references/Prompt - <Name>.md"
4. Follow the prompt's steps — use the shared scripts where the prompt calls for vault scans, daily note reads, or hub detection
5. Run temporal-track.sh if the prompt has a "Temporal Tracking" section (prevents redundant suggestions across runs)

Workflow Chains

Commands work together. Common chains:

Morning ritual: /context → /today Evening ritual: /close-day Weekly planning: /7plan → /drift Deep discovery: /emerge → /connect → /synthesize → /make Content pipeline: /ideas → /xarticle → /learned Strategy session: /money → /leverage → /evolve Vault maintenance: /map → /graph → /audit → /backlinks → /graduate

When a user asks for a chain (e.g., "plan my week"), run the commands in sequence, passing context forward.

Command Disambiguation

When it's unclear which command the user needs:

User wants...	Use	Not
Connections between 2 specific domains	/connect	/emerge (implicit patterns)
Ideas the vault implies but never states	/emerge	/ideas (generates new ones)
Weave 3+ threads into a narrative	/synthesize	/connect (only 2 domains)
Incompatible beliefs held now	/contradict	/challenge (pressure-tests)
How thinking changed over time	/trace	/contradict (simultaneous beliefs)
Vault structure narrative	/map	/graph (quantitative metrics)
Vault health + action plan	/audit	/map (no action plan)
Fix missing links	/backlinks	/map (only analyzes)
Score idea readiness to ship	/make	/ideas (generates, doesn't score)
Readiness x strategic leverage	/evolve	/make (readiness only)
What to write next for X	/xarticle	/ideas (all domains, not X-specific)
Daily note ideas → standalone	/graduate	/emerge (vault-wide, not daily notes)
Quantitative graph metrics	/graph	/map (qualitative narrative)

CLI Conventions

All commands use the obsidian CLI (lowercase). These conventions apply everywhere:

• format=json — Always use for machine-readable output when processing results
• counts format=json — On backlinks for quantitative hub detection
• jq — Slice/filter JSON output (e.g., | jq '.[0:20]')
• obsidian vault — Quick stats before any deep analysis
• obsidian recents — Catch recently modified non-daily notes
• obsidian tasks todo verbose format=json — Vault-embedded tasks
• obsidian wordcount file="X" — Measure thinking density (words, not just links)
• obsidian outline file="X" format=json — Structure before reading full content
• obsidian search:context — Use instead of search when you need matching lines
• obsidian backlinks file="X" counts format=json — Ranked connections, not just a list

Confidence Markers

Many commands reference confidence markers in context files:

• [solid] — High confidence, tested through experience
• [evolving] — Active development, changing
• [hypothesis] — Untested, speculative
• [questioning] — Under review, may change

Agents

For expensive operations, use the agents in agents/:

• vault-scanner (agents/vault-scanner.md) — Runs structural analysis in parallel. Use when a command needs vault scan + daily notes + context files simultaneously.
• thread-tracer (agents/thread-tracer.md) — Follows backlink chains 3+ hops deep. Use when Connect, Emerge, or 7plan need deep graph traversal.

Hooks

The hooks/hooks.json configures:

• SessionStart — Suggests relevant commands based on time of day and recent vault activity
• PostToolUse (Bash) — After obsidian CLI commands, suggests related OBC commands that could use the data