Decide Which Papers Deserve a Full Read

What it is for

A screening procedure for the moment you have imported forty documents and have time for five. It forces each summary to answer what was measured, on whom, and under what conditions your question is served, then sorts the pile into read, mine and drop. The judgement stays with you: it assumes only that the tool summarises and lets you search inside a document, never that it will follow a long instruction. It prevents the most common summarisation error, treating a well-written summary of a weak study as a citable finding. genei is still running and still billing on an auto-renewing trial, but the company's visible attention has moved to its newer product, so test it on one document before you trust a pile to it.

The skill Copy

Import the whole pile first, then screen it. Do not read in
import order - import order is arrival order, not importance.

Run one document through before you commit an afternoon to this.
Confirm you actually get a usable summary back for it. If you do
not, the pile is not the problem and no procedure fixes it.

For each document, read its summary and answer three questions in
your own words. If you cannot answer one, that is a finding, not
a failure:

  1. What did they actually measure or observe?
  2. On whom or what, and how many?
  3. What would have to be true for this to apply to my question?

When the summary is too thin to answer those, do not argue with
the tool in a longer question. Search inside the document for the
words that carry the answer - sample, participants, method,
outcome, control - and read those lines yourself. Write NOT STATED
beside any of the three you cannot fill. Two NOT STATEDs already
sort the paper.

Sort into three piles and write the pile name beside each title:

  READ  - answers your question directly, method is stated
  MINE  - you only want its reference list or one figure
  DROP  - wrong population, wrong outcome, or no method stated

Cap the READ pile at five. A sixth entry means something else
leaves.

The failure this prevents: a summary reads beautifully for a paper
whose method cannot carry your claim. Summaries flatten hedges -
"may be associated with" and "causes" come out looking the same -
and you end up citing a pilot study as if it were a trial.

Never cite from a summary. Before a document leaves the READ pile
and enters your draft, open the page the claim came from and read
the sentence yourself.

How to install it Copy

1. Open the tool and start a new conversation, project or template.
2. Paste the text below wherever the tool keeps standing instructions
   — usually called system prompt, instructions, or persona.
3. Save it so it applies to every run rather than a single message.

Did it work for you?

Where it came from