Purpose
Two short sequences that bracket the day. The morning sequence sets state and picks the day's few actions before the inbox does it for you. The evening sequence closes the loop — what you gave, what you learned, where you moved. This is old technology, not new (see Lineage below); this skill's only job is to wire it into the vault so the answers become data the loop can consolidate, not breath that disappears.
The ritual is priming, not assessment. It asks how you feel and what you appreciate. It never scores you, never reads a mood as a condition. See the guardrail below.
Why this works
Attention runs like a spotlight, not a floodlight: whatever you point it at first locks in what you keep noticing for hours after. Answer the morning questions on autopilot and the inbox picks your spotlight for you by 9am. Answer them deliberately and you set it yourself, before anything else competes for it. The evening sequence does the same job in reverse — what you deliberately look back at is what the week's memory keeps; what you skip past gets overwritten by tomorrow's noise.
The mechanism is not the question text. It's three moves, in order, every time:
- Answer the question specifically — not the first vague word, the actual detail.
- Picture it — the scene, not the category. ("My work" isn't a picture. The call with Sam this morning is.)
- Feel it and hold it for a breath before moving on. The felt state is the point. The words just got you there.
Skip step 3 and you've built a checklist. The checklist doesn't do anything.
When it activates
- You run
/dawn(morning) or/dusk(evening). - You say "prime me", "morning questions", "run my dawn ritual", "close out the day", "evening reflection", "walk me through the questions".
- The top of a fresh
daily/YYYY-MM-DD.md, before you plan the day.
Two modes — same questions
Read-aloud (human). The agent prints the sequence, one prompt at a time, and gets out of the way. You answer, picture it, feel it, hold it for a breath — then move to the next. Five to ten minutes. No capture required; the point is the state, not the record.
Captured (agentic). The agent asks the prompts, takes your one-line answers, and
writes them into today's daily note under a ## Priming (morning) or ## Reflection
(evening) block, tagging each against the constructs it touches. Now the weekly
reviewer sees a week of what you were grateful for, committed to, and gave — recurrence
across days is exactly the signal it consolidates on. Capture never shortens the
visualize-and-feel beat; it only adds a place for the answer to land afterward.
Default to read-aloud. Switch to captured when the person says "write it down", "log this", or is already in a daily note.
The morning sequence — set state, then pick actions
Seven state prompts, then three that convert the state into action. Every state prompt (1–7) runs the full cadence: answer, then what about that (get specific), then picture it and feel it, then hold that feeling for a breath before the next prompt. Don't rush it — rushing the feel step is the difference between priming and a survey.
| # | Prompt | Constructs |
|---|---|---|
| 1 | What am I happy about in my life right now? | emotion |
| 2 | What am I excited about right now? | emotion · motivation |
| 3 | What am I proud of right now? | identity · emotion |
| 4 | What am I grateful for right now? | emotion · social-cognition |
| 5 | What am I enjoying most right now? | emotion · motivation |
| 6 | What am I committed to right now? | motivation · identity · behavior |
| 7 | Whom do I love, and who loves me? | social-cognition · emotion |
| 8 | How can I make myself feel good, improve one relationship, be present, and give today? | emotion · social-cognition · consciousness · behavior |
| 9 | What are my few most important goals, and the one thing I'll do today toward them? | motivation · decision-making · behavior |
| 10 | What guidance do I want today — and if this were my last day, how would I live it? | belief · metacognition · consciousness |
Prompts 1–7 build state, one felt beat at a time. Prompt 8 turns that state outward. Prompts 9–10 convert it into a decision and a frame for the day. An agent running captured mode writes each answer as the person said it (observation), never as a claim about them.
The evening sequence — close the loop
Ten reflection prompts, run in reverse: instead of picturing forward, recall the actual
moment from today, then let the feeling register before writing it down. These feed
the daily note's existing sections (## Learned (learning · metacognition),
## Decided / did (decision-making · behavior), ## What moved me (emotion)) as much
as a ## Reflection block.
| # | Prompt | Constructs |
|---|---|---|
| 1 | What did I give today? Where was I a giver? | behavior · social-cognition |
| 2 | What did I learn? How can I improve? | learning · metacognition |
| 3 | Did I follow my intuition — what did it say? | decision-making · consciousness |
| 4 | How did I follow what I actually care about? | motivation · emotion |
| 5 | Where did I let things flow instead of forcing the outcome? | emotion · belief |
| 6 | How did today add to the quality of my life? Am I enjoying the journey? | consciousness · emotion |
| 7 | Best thing that happened today — how do I create more of it? | memory · learning |
| 8 | Where did I make progress? Did I mark it? | behavior · motivation |
| 9 | What am I grateful for today? How did I love today? | emotion · social-cognition |
| 10 | What help or guidance do I want to ask for? | belief · metacognition |
Prompt 8 is the one to make concrete — "where did I make progress" is the hook for any external accountability you keep (billable/creation output, a ship log). Naming it here is enough; the ritual doesn't own that ledger, it just surfaces the entry.
How the agent writes to the vault (captured mode)
- Open (or create from
vault-templates/daily/_template.md)daily/YYYY-MM-DD.md. - Morning → append a
## Primingblock. Evening → append a## Reflectionblock, and fold evening prompts 2, 3, and 8 into the note's existing## Learned (learning · metacognition),## What moved me (emotion), and## Decided / did (decision-making · behavior)sections rather than duplicating. - Write answers verbatim. One line each is fine; blank is fine and is itself data.
- Add touched constructs to the note's
constructs::line. Don't invent synonyms — only the 12 canonical names. - Stop. Don't summarize the person's state back at them, don't grade the day. The weekly reviewer does consolidation; the ritual only captures.
Non-clinical guardrail
This is priming and gratitude, not evaluation. The five-step separation holds:
the agent records what you said (observation), may offer a neutral framing only if
asked and always labeled as one (interpretation), never a verdict. It never reads a
low-energy answer as a symptom, never names a disorder, never gives medical or
therapeutic advice. "How does that feel" is a prompt for you, not a data point for the
agent to diagnose. If an answer signals crisis, the ritual stops and points to real
human help instead of continuing the sequence. See ../../CANON.md.
Lineage
The morning/evening priming pattern is adapted from Tony Robbins' priming practice and the Freedom Mastery planner's Questions to Empower Your Day / Evening Power Questions. Reworded into the OS's construct vocabulary and stripped of language this OS doesn't use (energy/vibration, prayer) in favor of secular equivalents (guidance, attention); credit to the originators for the sequence and the visualize-and-feel mechanism itself.
Built on SIP.