Daily Workflow Automation

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):

- 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):

- 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
• Formatting rules: 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:

cat .claude/commands/.extract-daily-content-tracker.json | jq '.processed_files | to_entries | last'

Reset tracker (force re-extraction):

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