06 · Digests
Morning desk digest
An overnight digest ranked by what can change your open risk, with a source grade on every item.
- Level
- Starter
- Cadence
- Each session open, covering a window you set
- Read
- 7 min
- Updated
Definition
A morning desk digest is a ranked list of overnight items that can change risk you hold or a decision you already planned. Each item has a time, a source, and a grade: primary, secondary, or commentary. Commentary never outranks a primary source. The digest does not tell you the trade.
Who it is for
Traders who start the session with open risk and do not want a feed of everything that happened.
Job to be done
Produce a short overnight digest ordered by impact on positions you already have, with a source grade on each item.
Human gate. You mark items as act, watch, or ignore. The digest does not place or suggest an order.
Inputs and outputs
Inputs
- Your open positions and resting orders, or a note that the book is flat
- The window, such as the last 12 hours
- Sources you accept as primary for those positions
- A max item count, small enough to read before the open
Outputs
- A ranked list, most relevant to your book first
- Source grade and time on every item
- A one-line "touches" note: which position or plan
- Items dropped because they do not touch the book
Step-by-step workflow
- 01Attach the bookPositions and working orders, even if the list is "flat except hedges". Ranking without this is entertainment.
- 02Set the window and the capA start time and a maximum number of items. The cap is what makes it a desk note.
- 03Grade every sourcePrimary: issuer, venue, protocol, filing. Secondary: reporting from those. Commentary: a person interpreting. Say which.
- 04Rank by your riskAn item that can change an open position outranks a larger story that does not touch you.
- 05Drop the restKeep a one-line count of dropped items so you know the filter worked. Do not sneak them back as "also".
- 06Human gateYou tag act, watch, or ignore. Act means you will look, not that an order exists.
Suggested tool types
These are categories of tools, not product picks. Crypto Agent Desk does not certify vendors and does not run an affiliate program.
| Type | Why it is here |
|---|---|
| Headline or RSS feeds you chose | The digest sorts a pile you defined. It should not free-associate the night. |
| Position snapshot | Ranking depends on what you hold. A generic market recap is a different product. |
| Chat model | Grades sources and drops irrelevant items. It should quote, not embellish. |
Risks and failure modes
Commentary dressed as news
A sharp take is not an event. If the grade is commentary, it cannot sit above a primary item that touches the same book.
Book-blind recap
Without positions, the model will rank by drama. Require the snapshot or label the digest as not book-aware.
Stale item
Overnight windows include articles that recap yesterday. Require a time. Drop what is outside the window.
Operator notes
The morning failure is reading the loudest item first. Rank by the book you have, and make commentary wear a label.
Item shape
Rank, time, grade, source, what happened in one sentence, which position or plan it touches, your tag left blank.
Prompt you can paste
Morning digest. Window: START to NOW. Cap: MAX_ITEMS.
Open risk: POSITIONS.
Items I collected: ITEMS with source and time.
Grade each source primary, secondary, or commentary.
Rank by relevance to open risk, not by drama.
Drop items that do not touch the book and report the drop count.
Do not suggest orders. Leave a tag field: act, watch, or ignore.
Done when
You can read it before the session and know which single item, if any, deserves a look at a live position.
FAQ
What is a source grade?
A label for how close the source is to the event. Primary beats secondary beats commentary when they conflict. The digest must show the grade, not hide it in tone.
How many items should survive?
Few enough to read before you touch the book. If you need a second page, the cap is wrong.
What if nothing touches the book?
Then the digest is "nothing overnight that changes open risk", which is a successful run.
Published . Updated . This is an operating playbook, not a recommendation to buy, sell, or hold any asset.