# MOST COMMON OPERATIONS
stamp fleeting # DEFAULT for ALL new info: 2025-11-12-F143000
stamp daily # ONLY for diary/journal: 2025-11-12
stamp project --check # ALWAYS check before creating project
arrowhead search fts "keyword" # Fast keyword search (~50-100ms)
arrowhead search hybrid "concept" # Comprehensive search (~400ms)
arrowhead graph backlinks "Note" # Find incoming links (~25ms)
things-cli anytime # View main task listGolden Rules:
- NEVER manually construct date-based filenames
- Use
stamp fleetingfor ALL new info (daily is ONLY for diary) - Prefer appending to existing notes
- Use [[wikilinks]] for people
- Check before creating (search first, create last)
This manual provides operational instructions for AI agents working with this Obsidian vault. The vault is a flat-hierarchy knowledge management system implementing [[P0450 Plaintext Commons]] - an experimental file-native thinking system with plain Markdown files as source of truth, Arrowhead indexing, and AI-assisted editing.
NEVER manually construct date-based or sequential filenames. The stamp tool is the ONLY authoritative source.
# Information capture (use these for ALL new info)
stamp fleeting # β 2025-11-12-F143000 (DEFAULT for all new info)
stamp voice # β 2025-11-12-VT143000 (voice transcripts)
stamp analog # β 2025-11-12-A1 (paper note photos)
# Journal entry (ONLY for personal diary)
stamp daily # β 2025-11-12 (diary/journal ONLY)
# Project operations
stamp project --check # Check next number FIRST
stamp project "Title" # β P0455 Title
# Review operations
stamp monthly # β 2025-11
stamp yearly # β 2025- Flat hierarchy mandatory - All notes in root (except Attachments/ and Leaves/)
- Never duplicate information - Link to existing notes instead
- Prefer appending to existing notes over creating new ones
- Use [[wikilinks]] for people and internal references only
- No manual date construction - Always use stamp CLI
- Frontmatter always comes first - YAML frontmatter must be at the very beginning of the file, followed by the H1 header
- Check existing notes with Arrowhead CLI before creating new ones
- Use
stamp fleetingfor new information capture (NOT daily - that's for diary only) - Append to existing project notes when appropriate
Decision tree:
ββ Know exact keywords? β arrowhead search fts (fastest ~50-100ms)
ββ Exploring concepts? β arrowhead search semantic (~400ms)
ββ Need both? β arrowhead search hybrid (~400ms)
ββ Need connections? β arrowhead graph context, graph backlinks (~25ms)
- Search existing content with Arrowhead CLI
- Read specific notes with Read tool
- Check connections with
arrowhead graph context "Note" - Use stamp CLI for new filename if needed
- Prefer Edit/append over Write for existing notes
| Command | Use Case | Speed | Output Formats |
|---|---|---|---|
arrowhead search fts "query" |
Known keywords, titles, metadata filters | β‘ ~50-100ms | human, ids, paths, json |
arrowhead search semantic "concept" |
Conceptual exploration, theme discovery | π ~400ms | human, ids, paths, json |
arrowhead search hybrid "topic" |
Comprehensive search when uncertain | π ~400ms | human, ids, paths, json |
arrowhead graph backlinks "Note" |
Find notes linking TO target | β‘ ~25ms | human, json |
arrowhead graph forward-links "Note" |
Find outgoing links from note | β‘ ~10ms | human, json |
arrowhead graph context "Note" |
Full graph view (backlinks + forward links) | β‘ ~35ms | human, json |
arrowhead graph orphans |
Find notes with no links | β‘ ~11ms | human, json |
arrowhead graph unresolved |
Find broken WikiLinks | β‘ ~12ms | human, json |
arrowhead notes read "Note ID" |
Get full content after search | β‘ Instant | - |
arrowhead notes list |
List all notes in vault | β‘ Fast | human, json |
Format Options:
--format human(default): Multi-column, readable output with scores and previews--format ids: Note IDs only, one per line (perfect for piping)--format paths: Absolute file paths, one per line (for xargs workflows)--json: Machine-readable JSON with full metadata
| Tool | Use Case | When to Use |
|---|---|---|
Read |
Read known files | You have exact path |
Edit |
Modify existing content | After Read, precise changes |
Write |
Create new files | Only when truly new |
Bash |
Run stamp CLI | Filename generation |
| Tool | Use Case | Details |
|---|---|---|
things-cli |
Task management | View/filter todos (read-only) |
open "things:///add..." |
Create tasks | URL scheme for task creation |
mcp__iMPC__* |
Calendar/contacts | Meeting scheduling, people info |
CRITICAL DISTINCTION:
- Fleeting notes = ALL new information capture (meetings, ideas, processing, research)
- Daily notes = STRICTLY personal diary/journal entries only
| Category | Purpose | Stamp Command | Example |
|---|---|---|---|
| fleeting | DEFAULT for all new info | stamp fleeting |
2025-11-12-F143000 |
| journal | Personal diary ONLY | stamp daily |
2025-11-12 |
| project | Project documentation | stamp project "Title" |
P0455 Title |
| evergreen | Refined permanent notes | Manual naming | File over app |
| person | People profiles | Manual naming | Lorenzo Cipriani |
| literature | Source material (verbatim) | Manual naming | The Creative Act |
- S#### Title - Photo collections (e.g.,
S0282 Out of Memory Vol 1) - JN### Title - Jinny Street Gallery exhibitions (e.g.,
JN001 Tokyo Flaneur)
CRITICAL: Literature notes must be verbatim copies. Never edit source content.
CRITICAL: All markdown files must follow this exact structure:
---
category: project/literature/evergreen/etc
status: active/completed/archive
created: YYYY-MM-DD
[additional fields as needed]
---
# Title of Note
Content begins here...WRONG (H1 before frontmatter):
# My Note Title
---
category: project
---CORRECT (Frontmatter first):
---
category: project
---
# My Note Titlestamp project --check # ALWAYS check before creating project
stamp analog --check # Check today's analog counter
stamp project --counter # View current project counter (454)
stamp analog --counter # View today's analog counter
stamp project --set 500 # Correct counter if neededConfiguration: Tokyo timezone, Project counter at P0454, Analog resets daily
Calendars: Toto + Ludo (shared), Toto (Private), Landscape, Birthdays, Japan Holidays, F1
Calendar Operations:
mcp__iMPC__calendars_list # List all calendars
mcp__iMPC__events_fetch # Get events with filtering
mcp__iMPC__events_create # Create new eventsContact Operations:
mcp__iMPC__contacts_me # Your personal info
mcp__iMPC__contacts_search # Find by name/phone/email
mcp__iMPC__contacts_create # Add new contactsMeeting Notes Workflow:
events_fetchβ Get meeting detailscontacts_searchβ Find attendee infostamp fleetingβ Create timestamped note for meeting- Add notes with [[person wikilinks]]
- Create follow-ups in Things (not iMPC)
Event Prefix Convention: T: (Toto), L: (Ludo), Jinny: (Gallery)
Key Commands:
things-cli anytime # PRIMARY: Main working list
things-cli today # HIGH PRIORITY: Today's urgent items
things-cli inbox # New tasks to triage
things-cli upcoming # Future scheduled
things-cli search "query" # Search tasks
things-cli --filter-project "Name" todos # Filter by project
things-cli --json anytime | jq # Parse as JSONCreating Tasks (URL Scheme - always to inbox):
open "things:///add?title=Task%20name"
open "things:///add?title=Task¬es=Context%20info"
open "things:///add?title=Task&when=today"Task Title Format: Project: specific task (e.g., Synapse: fix search bug)
WORKFLOW: Check anytime (primary) and today (priority) lists
YYYY-MM-DD: Brief description
β’ First change
β’ Second change
β’ Additional changes
Example: 2025-09-22: Daily notes and project updates
stamp fleetingβ Get timestamped filename for new info- Create note with [[person wikilinks]]
- Create follow-up tasks:
open "things:///add?title=Follow%20up%20with%20Person"
stamp project --checkβ Verify next number- Search existing:
arrowhead search fts "project term" --limit 10 - Append to existing P#### or create new with
stamp project "Title"
stamp voiceorstamp analogβ Get filename- Clean up transcription while preserving thought flow
- Add [[wikilinks]] to related notes
arrowhead search hybrid "topic" --limit 20β Comprehensive searcharrowhead graph context "Note"β Find connections- Append to existing evergreen or create new descriptive-named note
# Search commands (all support --limit, --format, --json)
arrowhead search fts "keyword" # Full-text search (~50-100ms)
arrowhead search semantic "concept" # Semantic vector search (~400ms)
arrowhead search hybrid "query" # Combined FTS + semantic (~400ms)
# Graph commands (all support --json)
arrowhead graph backlinks "Note Title" # Incoming links (~25ms)
arrowhead graph forward-links "Note" # Outgoing links (~10ms)
arrowhead graph context "Note" # Full graph view (~35ms)
arrowhead graph orphans # Notes with no links (~11ms)
arrowhead graph unresolved # Broken WikiLinks (~12ms)
# Note operations
arrowhead notes read "Note ID" # Read note content
arrowhead notes list # List all notes
arrowhead notes list --ids-only # Just note IDs
# Daemon management
arrowhead status # Live daemon status
arrowhead vault start # Start daemon
arrowhead vault stop # Stop daemonFTS search supports field:value syntax:
arrowhead search fts "category:project" # All project notes
arrowhead search fts "status:active" # Notes with status:active
arrowhead search fts "tag:photography" # Tagged notes
arrowhead search fts "photography category:project" # Combined filters# Human-readable (default)
arrowhead search fts "Arrowhead" --limit 3
# β Shows: title, score, BM25 rank, preview, match reason
# IDs only (for piping)
arrowhead search fts "category:project" --format ids
# β One note ID per line: P0461 Arrowhead CLI
# Paths only (for file operations)
arrowhead search semantic "photography" --format paths
# β Absolute paths: /Users/toto/Notes/totocaster/2025-02-10.md
# JSON (for parsing)
arrowhead search hybrid "query" --json | jq '.[] | .title'
# β Full metadata with scores, wikilinks, tags# Count matches
arrowhead search fts "photography" --format ids | wc -l
# Filter results
arrowhead search fts "P04" --format ids | grep "^P045"
# Process each match
arrowhead search semantic "review needed" --format paths | \
while read -r path; do open "$path"; done
# Find orphans from 2020
arrowhead graph orphans --json | \
jq -r '.orphan_notes[] | select(startswith("2020"))'
# Batch tag operation (conceptual)
arrowhead search fts "category:fleeting" --format ids | \
head -10 > notes-to-process.txt- FTS: ~50-100ms (instant feel)
- Semantic: ~400ms (sub-second)
- Hybrid: ~400ms (combines both)
- Graph ops: ~10-35ms (instant)
- Daemon: Auto-indexes on file changes
- Index: 1,534 notes indexed with 0 errors
Arrowhead shows why each result matched:
P0461 Arrowhead CLI 0.927 -18.80 P0461 Arrowhead CLI
Reason: Hybrid match (FTS 1.00, semantic 0.45)
Knowledge Management 0.441 N/A -
Reason: Semantic similarity 0.44
Declutering Camera Equipment 1.000 -6.19
Reason: Full-text match (rank -6.19)
- Hybrid: Shows both FTS and semantic scores
- Semantic: Shows similarity score (N/A for BM25)
- FTS: Shows BM25 rank (negative = better)
/totocaster/
βββ *.md # All notes (flat hierarchy)
βββ Attachments/ # Media files
βββ Leaves/ # Analog note photos
βββ .obsidian/ # Config
- [[P0450 Plaintext Commons]] - File-native thinking system being developed
- [[P0461 Arrowhead CLI]] - Fast Rust-based search/indexing CLI (production-ready)
- [[P0371 Project Humane Notebook System]] - Analog-digital notebook system
- [[Jinny Street Gallery]] - Art gallery operations
- [[The System]] - Core methodology documentation
- Owner: Toto Tvalavadze (see [[Toto Tvalavadze]] for full details)
- Company: SSKK (Simulacra and Simulation KK)
- Location: Tokyo (timezone for all timestamps)
- Languages: English primary, Georgian/Japanese acceptable
- Spelling: American English preferred
- [[The System]] - Core principles
- [[PHNS Workflow]] - Analog-digital bridge
- [[Values of The System Notetaking]] - Philosophy
Fleeting β Processing β Evergreen