Stats

---
name: stats
description: Show vault statistics and knowledge graph metrics. Provides a shareable snapshot of vault health, growth, and progress. Triggers on "/stats", "vault stats", "show metrics", "how big is my vault".
version: "1.0"
generated_from: "arscontexta-v1.6"
user-invocable: true
context: fork
model: sonnet
allowed-tools: Read, Grep, Glob, Bash
argument-hint: "[--share] — optional flag for compact shareable output"
---

## Runtime Configuration (Step 0 — before any processing)

Read these files to configure domain-specific behavior:

1. **`ops/derivation-manifest.md`** — vocabulary mapping
   - Use `vocabulary.notes` for the notes folder name
   - Use `vocabulary.note` / `vocabulary.note_plural` for note type references
   - Use `vocabulary.topic_map` / `vocabulary.topic_map_plural` for MOC references
   - Use `vocabulary.inbox` for the inbox folder name
   - Use `vocabulary.notes_collection` for semantic search collection name

2. **`ops/config.yaml`** — processing depth, automation settings

If no derivation file exists, use universal terms (notes, MOCs, etc.).

---

## EXECUTE NOW

**Target: $ARGUMENTS**

Parse immediately:
- If target contains `--share`: output compact shareable format after full stats
- If target is empty: output full stats display
- If target names a specific category (e.g., "health", "growth", "pipeline"): show only that category

**START NOW.** Collect metrics and present them.

---

## Philosophy

**Make the invisible visible.**

The knowledge graph grows silently. Without metrics, the user cannot tell whether their system is healthy, growing, stagnating, or fragmenting. /stats provides a snapshot that makes growth tangible — numbers that show progress, health indicators that catch problems, and trends that reveal trajectory.

The output should make the user feel informed, not overwhelmed. Metrics are evidence, not judgment. "12 orphans" is a fact. What to DO about it belongs to /graph or /{vocabulary.cmd_reflect}.

---

## Step 1: Collect Metrics

Gather all metrics. Run these checks in parallel where possible to minimize latency.

### 1a. Knowledge Graph Metrics

```bash
NOTES_DIR="{vocabulary.notes}"

# Note count (excluding MOCs)
TOTAL_FILES=$(ls -1 "$NOTES_DIR"/*.md 2>/dev/null | wc -l | tr -d ' ')
MOC_COUNT=$(grep -rl '^type: moc' "$NOTES_DIR"/*.md 2>/dev/null | wc -l | tr -d ' ')
NOTE_COUNT=$((TOTAL_FILES - MOC_COUNT))

# Connection count (all wiki links across notes/)
LINK_COUNT=$(grep -ohP '\[\[[^\]]+\]\]' "$NOTES_DIR"/*.md 2>/dev/null | wc -l | tr -d ' ')

# Average connections per note
if [[ "$NOTE_COUNT" -gt 0 ]]; then
  AVG_LINKS=$(echo "scale=1; $LINK_COUNT / $NOTE_COUNT" | bc)
else
  AVG_LINKS="0"
fi

# Topic count (unique values in topics: fields)
TOPIC_COUNT=$(grep -ohP '^\s*-\s*"\[\[([^\]]+)\]\]"' "$NOTES_DIR"/*.md 2>/dev/null | sort -u | wc -l | tr -d ' ')

# Link density
if [[ "$NOTE_COUNT" -gt 1 ]]; then
  POSSIBLE=$((NOTE_COUNT * (NOTE_COUNT - 1)))
  DENSITY=$(echo "scale=4; $LINK_COUNT / $POSSIBLE" | bc)
else
  DENSITY="N/A"
fi
```

### 1b. Health Metrics

```bash
# Orphan count (notes with zero incoming links)
ORPHAN_COUNT=0
for f in "$NOTES_DIR"/*.md; do
  NAME=$(basename "$f" .md)
  grep -q '^type: moc' "$f" 2>/dev/null && continue
  INCOMING=$(grep -rl "\[\[$NAME\]\]" "$NOTES_DIR"/ 2>/dev/null | grep -v "$f" | wc -l | tr -d ' ')
  [[ "$INCOMING" -eq 0 ]] && ORPHAN_COUNT=$((ORPHAN_COUNT + 1))
done

# Dangling link count
DANGLING_COUNT=$(grep -ohP '\[\[([^\]]+)\]\]' "$NOTES_DIR"/*.md 2>/dev/null | sort -u | while read -r link; do
  NAME=$(echo "$link" | sed 's/\[\[//;s/\]\]//')
  [[ ! -f "$NOTES_DIR/$NAME.md" ]] && echo "$NAME"
done | wc -l | tr -d ' ')

# Schema compliance (% of notes with required fields: description, topics)
MISSING_DESC=$(grep -rL '^description:' "$NOTES_DIR"/*.md 2>/dev/null | wc -l | tr -d ' ')
MISSING_TOPICS=$(grep -rL '^topics:' "$NOTES_DIR"/*.md 2>/dev/null | wc -l | tr -d ' ')
SCHEMA_ISSUES=$((MISSING_DESC + MISSING_TOPICS))
if [[ "$TOTAL_FILES" -gt 0 ]]; then
  # Notes with BOTH required fields
  COMPLIANT=$((TOTAL_FILES - MISSING_DESC))
  COMPLIANCE=$(echo "scale=0; $COMPLIANT * 100 / $TOTAL_FILES" | bc)
else
  COMPLIANCE="N/A"
fi

# MOC coverage
COVERED=0
for f in "$NOTES_DIR"/*.md; do
  NAME=$(basename "$f" .md)
  grep -q '^type: moc' "$f" 2>/dev/null && continue
  if grep -rl '^type: moc' "$NOTES_DIR"/*.md 2>/dev/null | xargs grep -l "\[\[$NAME\]\]" >/dev/null 2>&1; then
    COVERED=$((COVERED + 1))
  fi
done
if [[ "$NOTE_COUNT" -gt 0 ]]; then
  COVERAGE=$(echo "scale=0; $COVERED * 100 / $NOTE_COUNT" | bc)
else
  COVERAGE="N/A"
fi
```

### 1c. Pipeline Metrics

```bash
# Inbox items
INBOX_COUNT=$(find {vocabulary.inbox}/ -name "*.md" 2>/dev/null | wc -l | tr -d ' ')

# Queue pending (check both YAML and JSON formats)
QUEUE_FILE=""
if [[ -f "ops/queue/queue.yaml" ]]; then
  QUEUE_FILE="ops/queue/queue.yaml"
  QUEUE_PENDING=$(grep -c 'status: pending' "$QUEUE_FILE" 2>/dev/null || echo 0)
  QUEUE_DONE=$(grep -c 'status: done' "$QUEUE_FILE" 2>/dev/null || echo 0)
elif [[ -f "ops/queue/queue.json" ]]; then
  QUEUE_FILE="ops/queue/queue.json"
  QUEUE_PENDING=$(grep -c '"status": "pending"' "$QUEUE_FILE" 2>/dev/null || echo 0)
  QUEUE_DONE=$(grep -c '"status": "done"' "$QUEUE_FILE" 2>/dev/null || echo 0)
else
  QUEUE_PENDING=0
  QUEUE_DONE=0
fi

# Processed ratio (notes vs inbox)
TOTAL_CONTENT=$((NOTE_COUNT + INBOX_COUNT))
if [[ "$TOTAL_CONTENT" -gt 0 ]]; then
  PROCESSED_PCT=$(echo "scale=0; $NOTE_COUNT * 100 / $TOTAL_CONTENT" | bc)
else
  PROCESSED_PCT="N/A"
fi
```

### 1d. Growth Metrics

```bash
# This week's growth (notes with created: date within last 7 days)
WEEK_AGO=$(date -v-7d +%Y-%m-%d 2>/dev/null || date -d '7 days ago' +%Y-%m-%d 2>/dev/null)
if [[ -n "$WEEK_AGO" ]]; then
  THIS_WEEK_NOTES=$(grep -rl "^created: " "$NOTES_DIR"/*.md 2>/dev/null | while read -r f; do
    CREATED=$(grep '^created:' "$f" | head -1 | awk '{print $2}')
    [[ "$CREATED" > "$WEEK_AGO" || "$CREATED" == "$WEEK_AGO" ]] && echo "$f"
  done | wc -l | tr -d ' ')
else
  THIS_WEEK_NOTES="?"
fi

# This week's connections (approximate — count links in recently created notes)
if [[ "$THIS_WEEK_NOTES" -gt 0 && -n "$WEEK_AGO" ]]; then
  THIS_WEEK_LINKS=$(grep -rl "^created: " "$NOTES_DIR"/*.md 2>/dev/null | while read -r f; do
    CREATED=$(grep '^created:' "$f" | head -1 | awk '{print $2}')
    [[ "$CREATED" > "$WEEK_AGO" || "$CREATED" == "$WEEK_AGO" ]] && grep -oP '\[\[[^\]]+\]\]' "$f" 2>/dev/null
  done | wc -l | tr -d ' ')
else
  THIS_WEEK_LINKS="?"
fi
```

### 1e. System Metrics

```bash
# Self space
if [[ -d "self/" ]]; then
  SELF_FILES=$(find self/ -name "*.md" 2>/dev/null | wc -l | tr -d ' ')
  SELF_STATUS="enabled ($SELF_FILES files)"
else
  SELF_STATUS="disabled"
fi

# Methodology notes
METHODOLOGY_COUNT=$(ls -1 ops/methodology/*.md 2>/dev/null | wc -l | tr -d ' ')

# Observations pending
OBS_PENDING=$(grep -rl '^status: pending' ops/observations/ 2>/dev/null | wc -l | tr -d ' ')

# Tensions pending
TENSION_PENDING=$(grep -rl '^status: open\|^status: pending' ops/tensions/ 2>/dev/null | wc -l | tr -d ' ')

# Sessions captured
SESSION_COUNT=$(ls -1 ops/sessions/*.md 2>/dev/null | wc -l | tr -d ' ')
```

Adapt all directory names to domain vocabulary. Skip checks for directories that do not exist — report "N/A" instead of errors.

---

## Step 2: Format Output

### Full Output (default)

Generate a progress bar for the Processed metric:

```
Progress bar calculation:
  filled = PROCESSED_PCT / 5 (number of = characters out of 20)
  empty = 20 - filled
  bar = [===...   ] PCT%
```

```
--=={ stats }==--

  Knowledge Graph
  ===============
  {vocabulary.note_plural}:  [NOTE_COUNT]
  Connections:               [LINK_COUNT] (avg [AVG_LINKS] per {vocabulary.note})
  {vocabulary.topic_map_plural}:   [MOC_COUNT] (covering [COVERAGE]% of {vocabulary.note_plural})
  Topics:                    [TOPIC_COUNT]

  Health
  ======
  Orphans:      [ORPHAN_COUNT]
  Dangling:     [DANGLING_COUNT]
  Schema:       [COMPLIANCE]% compliant

  Pipeline
  ========
  Processed:    [==============      ] [PROCESSED_PCT]%
  Inbox:        [INBOX_COUNT] items
  Queue:        [QUEUE_PENDING] pending tasks

  Growth
  ======
  This week:    +[THIS_WEEK_NOTES] {vocabulary.note_plural}, +[THIS_WEEK_LINKS] connections
  Graph density: [DENSITY]

  System
  ======
  Self space:      [SELF_STATUS]
  Methodology:     [METHODOLOGY_COUNT] learned patterns
  Observations:    [OBS_PENDING] pending
  Tensions:        [TENSION_PENDING] open
  Sessions:        [SESSION_COUNT] captured

  Generated by Ars Contexta v1.6
```

### Interpretation Notes

After the stats block, add brief interpretation for any notable findings:

| Condition | Note |
|-----------|------|
| ORPHAN_COUNT > 0 | "[N] orphan {vocabulary.note_plural} — run `/graph health` for details" |
| DANGLING_COUNT > 0 | "[N] dangling links — run `/graph health` to identify broken links" |
| COMPLIANCE < 90 | "Schema compliance below 90% — some {vocabulary.note_plural} missing required fields" |
| OBS_PENDING >= 10 | "[N] pending observations — consider running /{vocabulary.rethink}" |
| TENSION_PENDING >= 5 | "[N] open tensions — consider running /{vocabulary.rethink}" |
| DENSITY < 0.02 | "Graph density is low — connections are thin. Run /{vocabulary.cmd_reflect} to strengthen the network" |
| PROCESSED_PCT < 50 | "More content in inbox than in {vocabulary.notes}/ — consider processing backlog" |
| THIS_WEEK_NOTES == 0 | "No new {vocabulary.note_plural} this week" |

Only show interpretation notes when conditions are notable. A healthy vault gets just the stats, no warnings.

---

## Step 3: Shareable Format (--share flag)

If invoked with `--share`, output a compact markdown block suitable for sharing on social media or in documentation:

```markdown
## My Knowledge Graph

- **[NOTE_COUNT]** {vocabulary.note_plural} with **[LINK_COUNT]** connections (avg [AVG_LINKS] per {vocabulary.note})
- **[MOC_COUNT]** {vocabulary.topic_map_plural} covering [COVERAGE]% of {vocabulary.note_plural}
- Schema compliance: [COMPLIANCE]%
- This week: +[THIS_WEEK_NOTES] {vocabulary.note_plural}, +[THIS_WEEK_LINKS] connections
- Graph density: [DENSITY]

*Built with [Ars Contexta](https://github.com/arscontexta) v1.6*
```

The shareable format:
- Omits health warnings (positive framing for sharing)
- Omits pipeline state (internal detail)
- Omits system metrics (internal detail)
- Includes only growth-positive metrics
- Always includes the Ars Contexta attribution line

---

## Step 4: Trend Analysis (when history exists)

If previous /stats runs are logged in `ops/stats-history.yaml` (or similar), compare current metrics against the last snapshot:

```
  Trend (vs last check):
    {vocabulary.note_plural}: [N] (+[delta] since [date])
    Connections:              [N] (+[delta])
    Density:                  [N] ([up/down/stable])
    Orphans:                  [N] ([improved/worsened/stable])
```

If no history exists, skip trend analysis. Do NOT create the history file — that is /health's responsibility.

---

## Edge Cases

### Empty Vault (0 notes)

Show zeros gracefully:
```
--=={ stats }==--

  Your knowledge graph is new. Start capturing to see it grow.

  Knowledge Graph
  ===============
  {vocabulary.note_plural}:  0
  Connections:               0
  {vocabulary.topic_map_plural}:   0
  Topics:                    0

  Generated by Ars Contexta v1.6
```

Do not show health, pipeline, growth, or system sections for an empty vault — they would all be zeros or N/A.

### No Queue System

Skip the Pipeline section entirely. Do not show an error.

### No Self Space

Show "disabled" for self space line. Do not show an error.

### No ops/derivation-manifest.md

Use universal vocabulary (notes, MOCs, etc.). All metrics work identically.

### Very Large Vault (500+ notes)

The orphan and MOC coverage checks may be slow for large vaults. If {vocabulary.notes}/ has >200 files:
1. Run orphan detection with a simpler heuristic (check only for presence in any MOC, not full backlink scan)
2. Note: "Metrics approximate for large vault. Run /graph health for precise analysis."

### Platform-Specific Date Commands

macOS uses `date -v-7d`, Linux uses `date -d '7 days ago'`. The script tries both. If neither works, report "?" for growth metrics instead of failing.