Daily Diary#

Generate structured diary entries by reviewing conversation history.

Workflow#

Diary generation follows these steps:

1. Review today's conversations → extract topics, achievements, decisions, insights
2. Generate draft → fill structured template, write in first person
3. Deliver to user → present draft with prompts for additions
4. Archive → save to ~/diary/YYYY/MM/YYYY-MM-DD.md

Step 1: Review Conversations#

Scan all conversations from today. Extract:

• Topics discussed — projects, domains, problems worked on
• Concrete achievements — tasks completed, features shipped, issues resolved
• Key decisions — choices made and their rationale
• Challenges — blockers, bugs, frustrations encountered
• Insights — ideas, lessons learned, perspectives shifted

Sensitive data filter — ALWAYS strip all credentials, keys, personal identity information, server addresses, and configuration file contents from extracted content before including in diary. Redact paths containing usernames to ~/. Content the user explicitly marked as confidential must also be excluded.

Step 2: Generate Draft#

Write in first person ("I"). Tone: warm, natural, like writing to future self.

Use this structure:

text

📔 YYYY-MM-DD Diary

☀️ Today's Summary
(One sentence capturing the day's main thread)

✅ Achievements
- Achievement 1
- Achievement 2

🧭 Key Decisions
- Decision: rationale and conclusion

💡 Reflections
(Insights or lessons extracted from conversations)

📋 Tomorrow's Plan
- [ ] Task 1
- [ ] Task 2

🎯 One-Line Mood
(Leave blank for user to fill)

🏷️ Tags
#tag1 #tag2 #tag3

Quality rules:

• Be specific with details, avoid generic summaries
• Preserve the user's original phrasing when it captures something well
• Extract 3-5 topic tags from content (domains like #AI, #product; types like #decision, #milestone; moods like #productive, #focused)
• Omit sections with no content rather than filling with filler
• Double-check that no sensitive data made it into the draft before delivery

Writing style:

• State facts directly. Drop filler phrases like "Additionally", "It is worth noting", "In this context"
• Vary sentence length. Short punchy sentences, then longer ones with detail. Mix it up, avoid mechanical repetition
• Do not inflate significance. "Finished security hardening" not "Completed a landmark security upgrade marking a new era for the system"
• Use concrete facts over vague attribution. "Deployed rate limiting" not "Took a series of critical measures"
• Have opinions and warmth. Diaries are personal — say "Hit a nasty bug today" or "Really happy with this solution"

Step 3: Deliver to User#

Append this footer to every diary draft:

text

---
💬 To complete your entry:
1. How was your mood today? (one sentence is fine)
2. Anything to add or change?
3. Reply ✅ to archive as-is

Step 4: Archive#

User responseActionAdds contentMerge additions, then saveReplies ✅ or "ok"Save immediatelyNo reply within 24hSave as draft (status: unreviewed)

Save path: ~/diary/YYYY/MM/YYYY-MM-DD.md

Append status line at end of file:

text

---
*Generated by [agent-name] at HH:MM | Status: confirmed/unreviewed*

Empty Day Handling#

When no conversations exist for today:

1. Briefly inform user: "No conversation records today."
2. Offer 2 guided prompts from the reference file. See quick-prompts.md.
3. If user responds, build entry from their input.
4. If no response, skip today's entry.

Guided Prompts#

When the user says they have nothing to write, or when conversation content is thin, use prompts from quick-prompts.md to draw out content.

Rules: pick at most 2-3 questions per session. Choose based on the day's conversation topics. Keep the tone light — journaling should never feel like homework.

Cron Setup#

Add to cron/jobs.json for automated daily triggers:

json

{
  "name": "daily-diary",
  "schedule": { "expr": "0 20 * * *", "kind": "cron", "tz": "Asia/Shanghai" },
  "sessionTarget": "isolated",
  "payload": {
    "kind": "agentTurn",
    "message": "It's evening diary time. Review today's conversations and generate a diary draft following the daily-diary skill workflow.",
    "timeoutSeconds": 180
  },
  "delivery": { "channel": "feishu", "mode": "announce", "to": "chat:<group-id>" }
}

Adjust tz, channel, and to for your setup.

Storage Structure#

text

~/diary/
├── YYYY/
│   └── MM/
│       └── YYYY-MM-DD.md
└── weekly/          (future: auto-generated weekly summaries)

Security#

• Filter all sensitive data (credentials, personal identity info, etc.) before writing. When uncertain, omit.
• Diary files are local-only. Never transmit beyond the configured delivery channel.
• May read user personality files to match writing style, but never copy raw content into entries.
• Never access system configuration or credential directories unrelated to diary generation.
• Only deliver to the conversation owner. Drafts remain drafts until user confirms.

Scope#

This skill ONLY:

• Reads conversation history for content extraction
• Creates and edits files within ~/diary/
• Delivers diary drafts through the user's configured channel

This skill NEVER:

• Modifies files outside the diary directory (reading personality files for tone matching is allowed)
• Sends diary content to external services or third parties
• Deletes diary entries without explicit user approval
• Stores any sensitive information in diary text
• Makes network requests beyond the configured delivery channel