Journal

Trigger Phrases

• "journal:"
• "reflect:"
• "today I"
• "feeling [emotion]"
• "grateful for"
• "on my mind:"

Inputs

The user provides a journal entry inline with the trigger phrase. The entry can be:

• A sentence or two: "journal: Had a breakthrough on the caching layer today"
• A mood expression: "feeling overwhelmed — too many threads open"
• A gratitude note: "grateful for Sarah's help debugging that race condition"
• A reflection: "reflect: I notice I default to building instead of asking for help"

Workflow

Step 1: Parse the Entry

Extract the journal content from the user's message.

• Strip the trigger prefix ("journal:", "reflect:", etc.)
• Preserve the user's exact words — do not rewrite, summarize, or polish
• If the entry is just a trigger phrase with no content (e.g., "journal:"), ask: "What's on your mind?"

Step 2: Detect Mood

Analyze the entry for mood signals. Assign one primary mood:

• If mood is ambiguous, default to Neutral
• Do not ask the user to confirm the mood — just detect it quietly

Step 3: Detect Tags

Assign 1-3 tags based on content:

Step 4: Check for Today's Daily Page

Search {{notes}} for today's daily page.

• If it exists, the journal entry will be linked to it
• If it doesn't exist, create the entry standalone (don't create a daily page just for a journal entry)

Step 5: Create the Journal Entry

Write the entry to {{notes}} with this structure:

Date: [today's date]
Mood: [detected mood]
Tags: [detected tags]

[User's journal entry, exactly as written]

• If today's daily page exists, also add a brief reference to the journal entry in the daily page under a "Journal" section
• Keep the reference minimal: just the mood and first line of the entry

Step 6: Confirm

Respond with a single line confirming the save:

Logged — [Mood] [Tags]. [brief acknowledgment]

Examples of confirmations:

• "Logged — Energized, Work. Nice breakthrough."
• "Logged — Overwhelmed, Work. That's a lot of threads."
• "Logged — Grateful, Personal. Noted."

Do NOT:

• Offer advice or coaching
• Ask follow-up questions (unless entry was empty)
• Restate the journal entry back to the user
• Add motivational commentary

The confirmation should feel like a nod, not a conversation.

Notes

• Journal entries are private by default. Never post them to {{chat}} or any shared space
• If the user writes multiple sentences, keep them as one entry — don't split into separate entries
• Mood detection is a convenience, not a judgment. Keep it neutral in tone
• If the {{notes}} integration is not configured, fall back to writing a local markdown file at ~/.ai-workflow/journal/[date].md

Examples

User: "journal: Had a breakthrough on the caching layer. The problem was stale invalidation keys — once I switched to content-based hashing, everything clicked." Mood: Energized Tags: Work, Learning Response: "Logged — Energized, Work + Learning. Good solve."

User: "feeling overwhelmed — three PRs in review, two meetings back to back, and a production alert I haven't triaged yet" Mood: Overwhelmed Tags: Work Response: "Logged — Overwhelmed, Work. Full plate today."

User: "grateful for Jordan pairing with me on that flaky test. Would've taken me hours solo." Mood: Grateful Tags: Gratitude, Work Response: "Logged — Grateful, Work. Noted."

User: "reflect: I keep volunteering for tasks outside my sprint scope. Need to get better at saying 'not this sprint' without guilt." Mood: Neutral Tags: Reflection, Work Response: "Logged — Reflection, Work."