Daily Workflow Automation

Original source / Raw SKILL.md

---
name: Daily Workflow Automation
description: Automatically processes daily notes when user mentions diary entries, daily summaries, or end-of-day workflows. Extracts insights, contexts, and categorizes content from "### 🧠 Notes" sections. Use when user discusses daily notes, asks "what did I work on today", mentions end-of-day, or talks about extracting diary content.
---

# Daily Workflow Automation

Auto-process daily notes: extract entries → classify → route to appropriate files → update trackers.

## When This Runs

Auto-triggers when user:
- Mentions "today's diary", "daily note", "extract content"
- Asks "what did I work on today?"
- Natural end-of-day: "done for today", "wrapping up", "extract today"
- References "### 🧠 Notes" section

## Quick Workflow

1. Scan daily notes in `00 - Daily/` for unprocessed files
2. Extract entries from `### 🧠 Notes` sections
3. Classify each entry (Diary/Insight/Context/Idea) with AI
4. Analyze relevance to 24 context areas (business, tech, health, etc.)
5. Route to primary + secondary destinations (>80% confidence)
6. Transform diary narratives → factual knowledge (for context files)
7. Present routing plan with reasoning
8. Get user approval (y/n/edit)
9. Execute routing with cross-links
10. Update tracker (`.extract-daily-content-tracker.json`)
11. **Smart memory detection**: Suggest Serena/Graphiti updates

## Entry Classification

### Diary Entries
- **Primary**: `04-resources/diary.md` (always)
- **Secondary**: `[area]/diary-[area].md` (if >80% confidence)
- **Format**: Preserve original narrative with date headers

### Insights
- **Primary**: `01-areas/p-dev/insights.md` (always)
- **Secondary**: Never to context files (insights ≠ context)

### Context Entries
- **Primary**: `04-resources/context.md` (always)
- **Secondary**: Matching `CLAUDE.md` files (if >80% confidence)
- **Format**: Factual bullet points (no dates, no narrative)

### Ideas
- **Primary**: `04-resources/ideas.md` (always)
- **Secondary**: Never to context files

## Context Areas (24 files)

**Business** (8):
- `business/mokai/CLAUDE.md`
- `business/mokhouse/CLAUDE.md`
- `business/accounting/CLAUDE.md`
- `business/crypto/CLAUDE.md`
- `business/soletrader/CLAUDE.md`
- `business/SMSF/CLAUDE.md`
- `business/safia/CLAUDE.md`
- `business/trust/CLAUDE.md`

**Personal Development** (3):
- `p-dev/learning/CLAUDE.md`
- `p-dev/mindset/CLAUDE.md`
- `p-dev/psychedelics/CLAUDE.md`

**Health & Fitness** (4):
- `health-fitness/CLAUDE.md`
- `health-fitness/diet/CLAUDE.md`
- `health-fitness/medical/CLAUDE.md`
- `health-fitness/gym/CLAUDE.md`

**Tech** (5):
- `tech/CLAUDE.md`
- `tech/ableton/CLAUDE.md`
- `tech/mac/CLAUDE.md`
- `tech/ai/CLAUDE.md`
- `tech/ai/research-ai/context-engineering/CLAUDE.md`

**Other** (3):
- `claude-code/CLAUDE.md`
- `harry/CLAUDE.md`
- `people/CLAUDE.md`

## Diary → Context Transformation

**Original diary entry**:
```
"Today I built a Supabase database for financial tracking for MOKAI. Took forever to set up auth but got it working. Will use for invoice management."
```

**Context file format** (`tech/CLAUDE.md`):
```markdown
- Supabase: Financial tracking database with auth
  *From:* [[04-resources/diary#25-10-19 - Sun]]
  *Related:* [[business/mokai/context-mokai]]
```

**Context file format** (`business/mokai/CLAUDE.md`):
```markdown
- Financial tracking system: Supabase database for invoice management
  *From:* [[04-resources/diary#25-10-19 - Sun]]
  *Related:* [[tech/context-tech]]
```

**Transformations applied**:
- ✅ Extract topically relevant facts
- ✅ Rephrase to third-person
- ✅ Remove: dates, pronouns, temporal markers, feelings
- ✅ Format: Bullet points with cross-links
- ✅ Custom formulation for each context area

## Smart Memory Detection

After extraction, detect opportunities to update memory systems:

### Serena Memory Detection
**Triggers**: Technical patterns, workflows, commands, automation
**Keywords**: `workflow`, `pattern`, `convention`, `script`, `command`, `automation`, `process`

**Example**:
```
Entry: "Created new workflow using Supabase triggers for invoice automation"

Detection:
  System: Serena
  Memory: system_patterns_and_guidelines
  Confidence: 0.87
  Update: "Invoice automation: Supabase triggers reduce manual data entry"
```

### Graphiti Detection (3 instances)
**Triggers**: Strategic decisions, business events, personal insights
**Keywords**: `decided`, `strategy`, `planning`, `client`, `meeting`, `learned`, `insight`, `discovered`

**Routing**:
- `business/mokai/*` or `business/mokhouse/*` → **graphiti-mokai**
- `business/accounting/*`, `crypto/*`, `SMSF/*` → **graphiti-finance**
- Everything else → **graphiti-personal**

**Example**:
```
Entry: "Discussed the future direction of MOKAI with Jack and Brie over dinner"

Detection:
  System: Graphiti
  Instance: graphiti-mokai
  Confidence: 0.85
  Episode: "MOKAI strategic planning session with Jack and Brie discussing future business direction and growth opportunities"
```

## File References

- Daily note template: [reference/templates.md](reference/templates.md)
- Formatting rules: [reference/formatting-rules.md](reference/formatting-rules.md)
- Extraction script: `scripts/extract_content.py <date>`

## MCP Integration

**Serena MCP** (preferred for file operations):
- `mcp__serena__list_dir` - Find daily notes
- `mcp__serena__search_for_pattern` - Extract ### 🧠 Notes sections
- `mcp__serena__find_file` - Locate context files
- `mcp__serena__write_memory` - Update Serena memories

**Graphiti MCP** (knowledge graph storage):
- `mcp__graphiti__add_memory` - Store strategic insights

**Use Serena for pattern matching** (faster than Read tool)

## Tracker System

**Location**: `.claude/commands/.extract-daily-content-tracker.json`

**Tracks**:
- Last scan timestamp
- Processed files (with modification times)
- Entries processed per file
- Routing decisions with confidence scores

**View tracker**:
```bash
cat .claude/commands/.extract-daily-content-tracker.json | jq '.processed_files | to_entries | last'
```

**Reset tracker** (force re-extraction):
```bash
rm .claude/commands/.extract-daily-content-tracker.json
```

## Critical Rules

- ✅ **Always preserve frontmatter** (date, event) in daily notes
- ✅ **Use Serena MCP** for pattern matching (faster than Read)
- ✅ **Deduplicate insights** before storing in Graphiti
- ✅ **Never delete original content** during extraction
- ✅ **Transform diary → factual** for context files
- ✅ **Preserve narrative** for diary files
- ✅ **Require >80% confidence** for secondary routing
- ✅ **Create cross-links** bidirectionally
- ✅ **Auto-create diary files** with proper frontmatter
- ✅ **Get user approval** before memory updates

## User Approval Flow

After extraction, user chooses:

**[a] Update all** - Update all suggested memories (confidence >0.80)

**[s] Selective** - Choose which memories to update:
```
Which memories to update? (comma-separated)

1. Serena (system_patterns_and_guidelines) - Confidence: 0.87
2. Graphiti-MOKAI (strategic planning) - Confidence: 0.85
3. Graphiti-Personal (health insight) - Confidence: 0.78

Your choice: 2
```

**[n] Skip** - No memory updates

**[l] Later** - Queue for end-of-day batch review

## Example Workflow

**User**: "What did I work on today?"

**Claude**:
1. Scans `00 - Daily/` for today's note
2. Extracts entries from `### 🧠 Notes`
3. Classifies each entry
4. Presents routing plan:

```
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
📋 Extraction Plan for [[25-10-21 - Mon]]
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

Entry 1: "MOKAI uses Indigenous procurement..."
  Type: Context (96% confidence)
    Reason: Factual business practice information

  Primary: 04-resources/context.md
  Secondary:
    ✓ business/mokai/CLAUDE.md (98%)
      → Direct MOKAI business strategy reference
    ✓ harry/CLAUDE.md (87%)
      → Personal Indigenous identity context

Entry 2: "Finished mixing Nintendo track..."
  Type: Diary (94% confidence)
    Reason: Daily activity log, project completion

  Primary: 04-resources/diary.md
  Secondary:
    ✓ business/mokhouse/diary-mokhouse.md (97%)
      → MOK HOUSE client project completion

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
Actions: [y]es / [n]o / [e]dit / [s]kip entry
```

5. User approves: `y`
6. Executes routing with cross-links
7. Updates tracker
8. Detects memory opportunities:

```
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
🧠 Memory System Detection
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

📊 Graphiti (MOKAI):
  Episode: "MOKAI Indigenous procurement strategy"
  Confidence: 0.96
  Reason: Strategic business positioning

Would you like to update memory systems?
  [a] Update all suggested memories
  [s] Selective (choose which to update)
  [n] Skip memory updates
  [l] Later (remind me at EOD)
```

## Troubleshooting

**Issue**: AI misclassifies entries
- **Solution**: Use edit mode to correct, AI learns from patterns

**Issue**: Too many secondary destinations
- **Solution**: Increase confidence threshold (currently 0.8)

**Issue**: Diary files not created
- **Solution**: Verify auto-creation logic, check file permissions

**Issue**: Missing cross-links
- **Solution**: Verify append operations, check for write errors

**Issue**: Memory detection missing opportunities
- **Solution**: Review detection keywords, adjust confidence threshold

**Issue**: Wrong Graphiti instance selected
- **Solution**: Manually specify instance in selective mode

## Related Commands

- Fallback: `/extract-daily-content` (manual trigger)
- Memory: `/update-serena-memory`
- Documentation: `/report:document-system`

## Success Criteria

- ✅ Correctly classify entry types (>90% accuracy)
- ✅ Route to appropriate areas with good confidence
- ✅ Show clear reasoning for routing decisions
- ✅ Allow easy manual override
- ✅ Create valid cross-links bidirectionally
- ✅ Auto-create diary files with proper frontmatter
- ✅ Track all routing decisions
- ✅ Handle incremental runs efficiently
- ✅ Maintain file integrity across all destinations
- ✅ Respect >80% confidence threshold for secondary routing


---

## Referenced Files

> The following files are referenced in this skill and included for context.

### reference/templates.md

```markdown
# Daily Note Template Reference

This document describes the structure and patterns used in daily notes for proper extraction and categorization.

## Template Location

Main template: [98-templates/daily-note.md](../../../98-templates/daily-note.md)

## Frontmatter Structure

```yaml
---
type: daily note
date: "YYYY-MM-DD"
day: Mon
month: Jan
year: YYYY
event: ""  # Optional: special events/occasions
---
```

## Standard Sections

### 1. Header
```markdown
# 🌤️ Day - DDth MMM YY
```

### 2. Tasks Section (Dataview)
```dataview
TASK
WHERE file.name = this.file.name
```

### 3. Events Section (Dataview)
```dataview
TABLE WITHOUT ID
  file.link as "Event",
  date as "Date",
  time as "Time",
  location as "Location"
FROM "00 - Daily"
WHERE type = "event" AND date >= date(today) AND date <= date(today) + dur(30 days)
SORT date ASC
LIMIT 10
```

### 4. Notes Section (PRIMARY EXTRACTION TARGET)
```markdown
### 🧠 Notes

[User's daily entries go here - this is what gets extracted and categorized]
```

**Extraction Pattern**:
- Scan all content between `### 🧠 Notes` header and next H3 heading (or EOF)
- Extract individual entries (paragraphs separated by blank lines)
- Classify each entry using AI
- Route to appropriate destination files

**Example entries**:
```markdown
### 🧠 Notes

Built a Supabase database for MOKAI financial tracking today. Authentication setup was tricky but got it working. Will use for invoice management system.

Had an insight about customer retention - need to focus on post-sale support rather than just acquisition. This could be a differentiator for MOK HOUSE.

Idea: Create automated workflow for extracting daily notes using MCP servers. Could classify entries by type and route to appropriate areas.
```

### 5. MOK HOUSE Projects Dashboard (Optional)
```dataview
TABLE WITHOUT ID
  file.link as "Project",
  status as "Status",
  priority as "Priority",
  client as "Client"
FROM "02-projects/mokhouse"
WHERE type = "project"
SORT priority DESC, status ASC
```

## Entry Types and Recognition Patterns

### Diary Entries
**Indicators**:
- First-person narrative ("I did...", "Today I...")
- Daily activities and events
- Project completion updates
- Meeting summaries
- Temporal markers (today, this morning, yesterday)

**Example**:
```markdown
Finished mixing the Nintendo track for the client. Took about 4 hours of tweaking the vocal chain. Client approved on first submission.
```

### Insights
**Indicators**:
- Learnings and realizations
- Patterns observed
- Reflections on experiences
- Strategic thinking
- Keywords: "learned", "realized", "insight", "noticed", "pattern"

**Example**:
```markdown
Realized that most client complaints come from unclear project scope. Need to implement detailed SOWs with acceptance criteria for every project.
```

### Context Entries
**Indicators**:
- Factual information
- Technical details
- How-to knowledge
- Configuration notes
- No emotional content
- Keywords: "setup", "configured", "installed", "works by"

**Example**:
```markdown
Supabase RLS policies work by filtering rows at the database level before they reach the application. More secure than application-level filtering.
```

### Ideas
**Indicators**:
- Future possibilities
- Feature concepts
- Business opportunities
- Process improvements
- Keywords: "could", "what if", "idea", "maybe", "potential"

**Example**:
```markdown
What if we created a Chrome extension that auto-captures Notion pages to the daily note? Could save 10 minutes per day.
```

## Nested Sections

Daily notes may include nested subsections:

```markdown
### 🧠 Notes

#### Deep Work Session
- Completed database schema design
- Implemented auth middleware

#### Client Calls
- Meeting with MOKAI prospect
- Discussed SOW for pentest engagement
```

**Extraction Rule**: Scan ALL H4, H5, H6 subsections within `### 🧠 Notes` recursively.

## Dataview Integration

Daily notes use Dataview for dynamic content:

### Tasks Query
Shows tasks associated with the daily note file.

### Events Query
Shows upcoming events from the next 30 days from the `00 - Daily/` folder where `type = "event"`.

### Projects Query (MOK HOUSE specific)
Shows active MOK HOUSE projects from `02-projects/mokhouse/` sorted by priority and status.

## Special Markers

### Tags
```markdown
#mokai #client-work #insight
```

Tags within entries help with additional routing context.

### Links
```markdown
Related to [[business/mokai/context-mokai]] and [[02-projects/mokhouse/nintendo-project]]
```

Existing links are preserved during extraction and inform secondary routing decisions.

### Priorities
```markdown
🔥 URGENT: Follow up with Supply Nation application
```

Emoji markers indicate priority and can influence routing confidence.

## Cross-Linking Format

When entries are routed to destination files, cross-links are added:

```markdown
*From:* [[04-resources/diary#25-10-21 - Mon]]
*Related:* [[business/mokai/context-mokai]]
```

This creates bidirectional navigation between daily notes and context/diary files.

## File Naming Convention

Daily note files follow this pattern:
```
🌤️ Day - DDth MMM YY.md
```

Examples:
- `🌤️ Mon - 21st Oct 25.md`
- `🌤️ Tue - 22nd Oct 25.md`
- `🌤️ Wed - 23rd Oct 25.md`

**Parsing Pattern**: Extract date from filename for dating routed entries.

## Location

All daily notes stored in: `00 - Daily/`

## Related Templates

- **Diary file template**: `04-resources/diary.md` (narrative format with date headers)
- **Context file template**: `04-resources/context.md` (factual bullets with cross-links)
- **Insights file template**: `01-areas/p-dev/insights.md` (reflective format)
- **Ideas file template**: `04-resources/ideas.md` (concept format)

```

### reference/formatting-rules.md

```markdown
# Formatting Rules for Daily Workflow

This document specifies exact formatting rules for transforming, routing, and storing extracted daily note entries.

## Entry Type Formatting

### Diary Entries

**Characteristics**:
- Preserve first-person narrative voice
- Include date headers
- Maintain temporal context
- Keep emotional/subjective content

**Primary Destination**: `04-resources/diary.md`

**Format**:
```markdown
## YYYY-MM-DD - Day

[Narrative entry preserved exactly as written, including "I", "today", feelings, etc.]
```

**Secondary Destinations**: `[area]/diary-[area].md` (if confidence >80%)

**Format** (area-specific diary files):
```markdown
## YYYY-MM-DD - Day

[Narrative entry preserved, but filtered to area-relevant content only]

*From:* [[00 - Daily/🌤️ Day - DDth MMM YY#🧠 Notes]]
```

**Example**:
```markdown
## 2025-10-21 - Mon

Finished mixing the Nintendo track today. Took about 4 hours of vocal chain tweaking but client approved on first submission. Feeling good about the workflow improvements I made last month - they're really paying off now.

*From:* [[00 - Daily/🌤️ Mon - 21st Oct 25#🧠 Notes]]
```

---

### Context Entries

**Characteristics**:
- Transform to third-person factual statements
- Remove dates, pronouns, temporal markers
- Remove feelings and subjective assessments
- Extract topically relevant facts only
- Rephrase for each context area (different formulation per area)

**Primary Destination**: `04-resources/context.md`

**Format**:
```markdown
- [Topic]: [Factual statement]
  *From:* [[00 - Daily/🌤️ Day - DDth MMM YY#🧠 Notes]]
```

**Secondary Destinations**: Matching `CLAUDE.md` files in specific areas (if confidence >80%)

**Format** (area-specific context files):
```markdown
- [Topic]: [Factual statement custom-formulated for this area]
  *From:* [[00 - Daily/🌤️ Day - DDth MMM YY#🧠 Notes]]
  *Related:* [[path/to/related/context-file]]
```

**Transformation Rules**:
1. **Remove pronouns**: "I built" → "Supabase database built"
2. **Remove temporal markers**: "today", "this morning", "yesterday"
3. **Remove feelings**: "Took forever" → (omit), "feeling good" → (omit)
4. **Extract facts**: "Will use for invoice management" → "for invoice management"
5. **Topic-first**: Lead with the main subject
6. **Custom formulation**: Phrase differently for each area to reflect that area's perspective

**Example Transformation**:

**Original Diary Entry**:
```markdown
Today I built a Supabase database for financial tracking for MOKAI. Took forever to set up auth but got it working. Will use for invoice management.
```

**Primary Context** (`04-resources/context.md`):
```markdown
- Supabase: Financial tracking database with auth, for invoice management
  *From:* [[00 - Daily/🌤️ Mon - 21st Oct 25#🧠 Notes]]
```

**Secondary Context** (`tech/CLAUDE.md`):
```markdown
- Supabase: Financial tracking database with auth
  *From:* [[00 - Daily/🌤️ Mon - 21st Oct 25#🧠 Notes]]
  *Related:* [[business/mokai/context-mokai]]
```

**Secondary Context** (`business/mokai/CLAUDE.md`):
```markdown
- Financial tracking system: Supabase database for invoice management
  *From:* [[00 - Daily/🌤️ Mon - 21st Oct 25#🧠 Notes]]
  *Related:* [[tech/context-tech]]
```

---

### Insights

**Characteristics**:
- Preserve reflective/analytical tone
- Keep first-person if it enhances clarity
- Include date context if relevant to understanding
- Maintain strategic/learning focus

**Primary Destination**: `01-areas/p-dev/insights.md`

**Format**:
```markdown
## YYYY-MM-DD

[Insight preserved with minimal editing, maintaining analytical depth]

*From:* [[00 - Daily/🌤️ Day - DDth MMM YY#🧠 Notes]]
```

**Secondary Destinations**: NEVER route insights to context files (insights ≠ context)

**Example**:
```markdown
## 2025-10-21

Most client complaints stem from unclear project scope. Implementing detailed SOWs with acceptance criteria for every project could prevent 80% of scope creep issues.

*From:* [[00 - Daily/🌤️ Mon - 21st Oct 25#🧠 Notes]]
```

---

### Ideas

**Characteristics**:
- Preserve speculative/creative tone
- Keep future-oriented language ("could", "what if")
- Maintain brainstorming format
- No need for dates unless time-sensitive

**Primary Destination**: `04-resources/ideas.md`

**Format**:
```markdown
- [Idea statement]
  *From:* [[00 - Daily/🌤️ Day - DDth MMM YY#🧠 Notes]]
```

**Secondary Destinations**: NEVER route ideas to context files

**Example**:
```markdown
- Chrome extension that auto-captures Notion pages to daily note - could save 10 minutes per day
  *From:* [[00 - Daily/🌤️ Mon - 21st Oct 25#🧠 Notes]]
```

---

## Cross-Linking Format

All routed entries MUST include cross-links back to the source daily note.

### Link to Daily Note
```markdown
*From:* [[00 - Daily/🌤️ Day - DDth MMM YY#🧠 Notes]]
```

**Format Components**:
- `00 - Daily/` - Folder path
- `🌤️ Day - DDth MMM YY` - Exact filename
- `#🧠 Notes` - Section anchor (optional but recommended)

### Link to Related Context Files
```markdown
*Related:* [[path/to/context-file]]
```

**When to add Related links**:
- When an entry routes to multiple context areas
- Add bidirectional links between related areas
- Limit to 1-3 most relevant related files

**Example**:
```markdown
- Supabase: Financial tracking database
  *From:* [[00 - Daily/🌤️ Mon - 21st Oct 25#🧠 Notes]]
  *Related:* [[business/mokai/context-mokai]], [[tech/context-tech]]
```

---

## Auto-Creation Rules

### Diary Files

If area-specific diary file doesn't exist, create it with frontmatter:

```markdown
---
date created: Day, MM DDth YY, HH:MM:SS am/pm
date modified: Day, MM DDth YY, HH:MM:SS am/pm
---

# [Area Name] Diary

## YYYY-MM-DD - Day

[First entry]

*From:* [[00 - Daily/🌤️ Day - DDth MMM YY#🧠 Notes]]
```

**Locations**:
- `business/mokai/diary-mokai.md`
- `business/mokhouse/diary-mokhouse.md`
- `01-areas/p-dev/learning/diary-learning.md`
- etc.

### Context Files

Context files should already exist. If missing, create with:

```markdown
---
date created: Day, MM DDth YY, HH:MM:SS am/pm
date modified: Day, MM DDth YY, HH:MM:SS am/pm
---

# [Area Name] Context

- [First context entry]
  *From:* [[00 - Daily/🌤️ Day - DDth MMM YY#🧠 Notes]]
```

---

## Frontmatter Rules

### Required Frontmatter (for auto-created files)
```yaml
---
date created: Day, MM DDth YY, HH:MM:SS am/pm
date modified: Day, MM DDth YY, HH:MM:SS am/pm
---
```

**Format**:
- Day: `Mon`, `Tue`, `Wed`, etc.
- MM: `10`, `01`, `12`, etc.
- DDth: `21st`, `02nd`, `13th`, etc.
- YY: `25`, `26`, etc.
- Time: `6:24:50 am` or `3:15:30 pm`

### Never Modify Frontmatter (for existing files)
- Preserve existing frontmatter exactly
- Only update `date modified` field on edits (if it exists)

---

## Deduplication Rules

Before appending to destination files, check for duplicates:

### Fuzzy Match Threshold: 80%

Use Levenshtein distance or similar algorithm to detect near-duplicates.

### What to Check
- **Diary entries**: Check last 30 entries in the file
- **Context entries**: Check entire file (usually smaller)
- **Insights**: Check last 50 entries
- **Ideas**: Check entire file

### Duplicate Detection Examples

**Original Entry**:
```markdown
Supabase: Financial tracking database with auth
```

**Existing Entry (DUPLICATE)**:
```markdown
Supabase: Financial tracking DB with authentication
```

**Similarity**: 85% → Skip appending, log as duplicate

**Existing Entry (NOT DUPLICATE)**:
```markdown
Supabase: Project management database
```

**Similarity**: 60% → Append as new entry

---

## Append vs Insert Rules

### Always Append (Never Prepend)
Append new entries to the END of destination files to maintain chronological order.

### Section Placement

**Diary Files**:
```markdown
## 2025-10-20 - Sun
[Older entry]

## 2025-10-21 - Mon
[Newer entry] ← APPEND HERE
```

**Context Files**:
```markdown
- Old context entry
  *From:* [[old-daily-note]]

- New context entry ← APPEND HERE
  *From:* [[new-daily-note]]
```

**Insights Files**:
```markdown
## 2025-10-20
[Older insight]

## 2025-10-21 ← APPEND HERE
[Newer insight]
```

**Ideas Files**:
```markdown
- Old idea
  *From:* [[old-daily-note]]

- New idea ← APPEND HERE
  *From:* [[new-daily-note]]
```

---

## Special Characters and Escaping

### Preserve Special Characters
- Keep emojis: `🔥`, `📊`, `✅`
- Keep formatting: `**bold**`, `*italic*`, `` `code` ``
- Keep lists: `-`, `*`, numbered lists

### Escape Only When Necessary
- Markdown link brackets: Already handled by `[[wikilink]]` format
- Code blocks: Use triple backticks when code snippets are in entries

**Example**:
```markdown
- Configured Supabase RLS with this policy:
  ```sql
  CREATE POLICY "user_access" ON invoices FOR SELECT USING (auth.uid() = user_id);
  ```
  *From:* [[00 - Daily/🌤️ Mon - 21st Oct 25#🧠 Notes]]
```

---

## Validation Checklist

Before marking extraction as complete:

- [ ] All entries classified correctly (Diary/Insight/Context/Idea)
- [ ] Confidence scores >80% for secondary routing
- [ ] Diary entries preserve narrative tone
- [ ] Context entries transformed to factual statements
- [ ] All entries have `*From:*` links
- [ ] Related context files have bidirectional links
- [ ] No duplicate entries appended
- [ ] Frontmatter preserved on existing files
- [ ] New diary files created with proper frontmatter
- [ ] Tracker updated with processed files (but NOT today's file)
- [ ] User approved routing decisions

---

## Error Handling

### Missing Files
- **Diary/Context/Insights/Ideas base files missing**: Error and halt (these should always exist)
- **Area-specific files missing**: Auto-create with frontmatter

### Malformed Entries
- **No extractable content**: Log and skip, don't error
- **Ambiguous classification**: Default to Diary, log for manual review
- **Very short entries (<10 words)**: Classify but flag for review

### Write Failures
- **Permission errors**: Error and halt
- **Concurrent edit conflicts**: Retry once, then error
- **Disk full**: Error and halt

---

## File Integrity Rules

### Never Delete Original Content
- Daily note `### 🧠 Notes` section remains untouched after extraction
- Extraction is a **copy operation**, not a move operation

### Atomic Operations
- Write to temporary file first
- Verify write success
- Rename to final destination
- Prevents partial writes

### Tracker Updates
- Update tracker ONLY after ALL entries successfully routed
- Mark files as processed ONLY if today is PAST (allow re-processing today's file)
- Store confidence scores for audit trail

```