The Prompt I Use to Generate an Obsidian Friendly Log With Backlinks, Tags and Tasks
Every night, shortly before 12:00 AM, an automation I set up on my Mac opens the ChatGPT app and send this prompt:
Generate my Obsidian AI Daily Log for the last 24 hours.
Requirements:
- Sort everything in true chronological order with timestamps.
- Use one-line summaries for each discussion.
- Structure it as a technical journal, not a chat summary.
- Include these sections:
1) Timeline
2) Signals & Insights (what I learned or realized)
3) Friction & Problems (what broke, confused, or slowed things down)
4) Decisions (explicit or implicit choices made)
5) Open Questions (unresolved issues)
6) Next Actions (concrete follow-ups)
7) Backlinks (Obsidian-style [[topics]])
8) Tags (#macos #automation #clawdbot #health etc.)
- Use Obsidian-friendly Markdown.
- “At the end, include the full log again inside a plain-text block between ===BEGIN OBSIDIAN DAILY LOG=== and ===END OBSIDIAN DAILY LOG=== so I can save it to a file via automation. Do not use any clickable links for this block.”
When I wake up in the morning, there is a note in a folder in my vault, named using a repeatable convention so that I can create links each day in my daily note using a template with a variable.
An example note looks like this
## 🤖 AI Daily Log (23:52)
Obsidian AI Daily Log — Last 24 Hours
Timeline
- 01:02 — Generated Obsidian AI Daily Log for January 30 with full technical journal structure and export block.
- 02:44 — Investigated how the macOS utility LeaderKey works and explored command-style use cases beyond apps and URLs.
- 03:36 — Asked whether Calibre maintains logs and explored recommended settings and behavior of the Job Spy plugin.
- 04:44 — Requested a detailed narrative overview of Homeland Season 2, Episode 5.
- 07:05 — Analyzed the macOS security command sudo spctl --master-disable and its system-wide implications.
- 08:02 — Evaluated new eyeglass prescription with emphasis on computer-centric use and identified reputable online vendors.
Signals & Insights
- LeaderKey functions best as a command dispatcher layered atop existing automation.
- Calibre logging is present but distributed and inconsistent.
- Disabling Gatekeeper globally is risky and best reserved for short-term diagnostics.
- Computer-focused lenses align better with actual usage patterns.
- Homeland continues deliberate narrative destabilization in S2E5.
Friction & Problems
- Lack of centralized logging in Calibre complicates troubleshooting.
- Sparse advanced documentation for LeaderKey.
- macOS security controls encourage overcorrection due to poor visibility.
Decisions
- Use LeaderKey selectively as automation glue.
- Avoid permanent Gatekeeper disablement.
- Choose computer-optimized eyewear.
- Maintain AI logs as technical records.
Open Questions
- Best practices for persistent Calibre job logging?
- Community-standard LeaderKey command patterns?
- Safer macOS security toggling workflows?
Next Actions
- Prototype LeaderKey + shell/KM workflows.
- Capture Calibre debug logs during test jobs.
- Write a Gatekeeper troubleshooting checklist.
- Order glasses optimized for screen work.
Backlinks
[[Obsidian AI Daily Log]] [[macOS Security]] [[LeaderKey]] [[Calibre]] [[Automation]] [[Homeland]] [[Health]]
Tags
#tech/macOS #tech/automation #apps/leaderkey #apps/calibre #tech/security #obsidian #entertainment/television #health #vision