ravila4/claude-adhd-skills/daily-journal
Claude Code and agent skill by ravila4/claude-adhd-skills
- community
Checking that the SKILL.md file is still in the repository.
Draft, organize, or update development journal entries. Use when user requests daily journal generation, work review, standup preparation, or asks "what did I work on today/this week?"
Install
npx vskill@latest install ravila4/claude-adhd-skills/daily-journalSKILL.md
Daily Journal
An interactive, conversational skill for creating meaningful daily journal entries. Conversation first, data second.
Table of Contents
- Philosophy
- Flow Diagram
- Time-Aware Interaction
- The Conversation
- Using Data as Memory Aid
- Markdown Conventions
- Template Structure
- CLI Reference
Philosophy
Commits are not what matters. They're often mechanical, LLM-generated, or busywork. A journal should capture:
- What you actually focused on
- How you felt about the work
- What's still open in your mind
- The human experience of the day
Conversation first, data second. Ask the user about their day. Use collected data (git, sessions) only as a memory jogger for days when they can't remember what they did.
Context matters. Weekends are different from workdays. Mornings are different from evenings. Hobby projects are valid. Tooling optimization is valid.
Flow Diagram
Time-Aware Interaction
Check the current time before starting. Different times trigger different modes:
| Time | Mode | Primary Questions | |
|---|---|---|---|
| Before noon | Planning | What's your focus? Carryover from yesterday? | |
| Afternoon (12-5pm) | Mid-day check | How's the day going? Stuck on anything? | |
| Evening (after 5pm) | Reflection | What did you work on? How did it go? | |
| Weekend | Lighter touch | What did you tinker with? Any hobby wins? |
How to check:
Use this to:
- Determine mode (morning/afternoon/evening)
- Check if weekend (Saturday/Sunday) - add
weekendtag, lighter expectations - Frame questions appropriately
The Conversation
Morning Mode (Planning)
- Read yesterday's entry if it exists
- Show yesterday's "Tomorrow's focus", "Open questions"
- Ask:
- "What's your main focus for today?"
- "Any carryover from yesterday you want to tackle?"
- "Anything blocking you?"
- Create entry with intentions (can be updated in evening)
Evening Mode (Reflection)
- Start with open questions - don't assume anything from data:
- "What did you work on today?"
- "How did it go? What felt good or frustrating?"
- "Any wins? Any blockers?"
- "What's still on your mind?"
- Offer data as memory aid (only after user has shared):
- "I see you also had activity in [project] - worth mentioning?"
- "Looks like you touched [files] - anything notable there?"
- User can say "no, that was just mechanical" and it gets skipped
- Ask about tomorrow:
- "What's your focus for tomorrow?"
- "Any open questions to sleep on?"
Weekend Mode
- Lighter expectations, different framing
- "What did you tinker with today?"
- "Any hobby wins?"
- "Did you rest?" (rest is valid!)
- Add
weekendtag automatically
Using Data as Memory Aid
The collect_day_data.py script is a secondary tool, not the primary input.
When to use it:
- User says "I can't remember what I did"
- After user has shared, to jog memory for forgotten items
- Never as the source of truth for what mattered
How to present it:
Running the script:
Markdown Conventions
Clearly differentiate personal voice from LLM context using Obsidian-compatible markdown:
Personal Reflections (User's Voice)
Use blockquotes for the user's own words and feelings:
Context/Background (LLM-Provided)
Use regular text or callouts for factual context:
Technical Notes (If Needed)
Use collapsible sections for technical details that shouldn't dominate:
Callout Types (Obsidian)
| Callout | Role | Example titles | |
|---|---|---|---|
[!info] | Context, continuity, collapsible references | "Context", "Reference URLs" | |
[!success] | Wins, milestones, progress | "Win", "Big day", "Progress" | |
[!warning] | Concerns, friction, blockers | "Blocker", "Growing Pain" | |
[!question] | Open loops, things still thinking about | "Still thinking about..." | |
[!bug] | Bugs discovered | "Script Bug Found" | |
[!note] | Observations, non-actionable insights | "Observation: ..." |
Use the collapsible variant ([!info]-) for bulky reference material (URLs, log snippets).
See callout-conventions.md for detailed examples and placement guidance.
Template Structure
The final journal should feel personal, not mechanical:
CLI Reference
| Flag | Description | |
|---|---|---|
--date YYYY-MM-DD | Target date (default: today) | |
--quick | Skip interactive questions, just use data (not recommended) |
Editing Past Entries
Never delete historical context. Journal entries are a record of thinking over time - even wrong thinking has value.
- Cross out outdated items with strikethrough:
~~Outdated info~~ - Add a brief explanation when crossing out:
~~Thought X caused the bug~~ Turned out to be a race condition - Only delete if something is factually incorrect (e.g., wrong date, wrong project name)
- Completed items, debunked theories, abandoned plans, tasks no longer relevant -> strike through, don't delete
Examples:
Day Footer
Every entry ends with a footer: a horizontal rule followed by an italicized one-liner that captures the overall character of the day. This goes after the last content section and before the Related links.
Format: *{Day character} - {One-line takeaway}*
The "day character" is a brief descriptor of the day's shape -- how long, what kind of energy. The takeaway summarizes the arc.
Guidelines:
- Always include when closing off an entry at end of day
- Day character can mention: duration (short/full/long), time of day (morning/evening/overnight), energy (productive/scattered/chill)
- Takeaway should be conversational, not a formal summary
- Keep it to one line
Anti-Patterns
Don't:
- Lead with commits or session counts
- Generate a "report" from data
- Assume commits = what mattered
- Write in third person about the user
- Fill in feelings or reflections without asking
- Delete outdated entries, completed items, or debunked theories (strike them out instead)
Do:
- Ask open questions first
- Let user describe in their own words
- Use data only as memory jogger
- Preserve the user's voice in the final output
- Respect that rest and hobby time are valid
Security
- Security scanPassed, 95 of 100Pattern checks on the skill's files.
- Model reviewPassed
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
- SKILL.md in the repositoryCheckingskills/daily-journal/SKILL.md
Popularity
- 7-day trend score
- 0
Not refreshed yet.
Works with
Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.
- Claude Code
- Cursor
- GitHub Copilot
- Windsurf
- Codex
- Gemini CLI
- Cline
- Amp
- Roo Code
- Goose
- Continue
- Kilo Code
- Junie
- Trae
- OpenHands
- Qwen Code
- AI Agent
- Tabnine Chat
- Sourcegraph Cody
- Amazon Q
- Firebase Genkit
- Sweep
- Devin
- Phind
- Mentat
- Claude Desktop
- Aider
- GPT Pilot
- Plandex
- bolt.new
- v0
- Replit Agent
- Anse
- TypingMind
- BetterChatGPT
- LibreChat
- Warp AI
- Pieces
- Fig AI
Embed badge
[](https://verified-skill.com/skills/ravila4/claude-adhd-skills/daily-journal)