Read the Paper, Keep the Citations
Summarise technical papers without losing what claim came from where.
A summary that drops citations is worse than no summary, because you will repeat its claims in a client meeting with nothing behind them.
That is the gap this drop closes. It is a data & analysis skill built for Kimi, and it takes about ten minutes to set up the first time. After that it runs in under a minute.
Who should use it
Consultants and analysts who must cite what they claim.
How it works
The skill file does five things, in this order.
1. Load the full papers. Methods and limitations sections included, not just abstracts.
2. Tie every claim to a location. Section and page, so you can check it in the meeting.
3. Separate finding from speculation. Authors' own discussion sections blur this constantly.
4. Extract the limitations. Sample size, timeframe, funding source, and generalisability.
5. Write the citable version. Each sentence carrying its own reference.
What comes back
A claim-by-claim summary with locations, limitations, and confidence levels.
The mistake to avoid
Always ask for the funding source and the limitations section explicitly. Both are usually near the end and are the first thing a sceptical client will ask about.
Running it
Paste the prompt block into a new Kimi chat and attach your files in the same message. Kimi's long context means you should upload everything at once rather than in batches, splitting the input is what causes it to lose cross-references between documents.
Where this fits
On its own, one skill saves an hour a week. The compounding happens when three or four of them run in sequence on the same input, the same transcript that produces a scope of work also produces the follow-up email and the project brief. That is the point at which it stops being a prompt and starts being an internal tool. If you want that wired into the systems your team already uses, that is the work 67 Digital does.
In the file
Summarise these papers so every claim stays traceable.
For each paper:
1. WHAT THEY DID, method in three sentences, plain language.
2. FINDINGS, bullet list. Every bullet ends with (Section X, p.Y). Never
merge findings from different papers into one bullet.
3. WHAT THEY DID NOT FIND, results explicitly reported as null or
inconclusive. These are usually omitted from press coverage.
4. LIMITATIONS, sample, timeframe, funding source, conflicts, and stated
generalisability limits.
5. CONFIDENCE. HIGH / MEDIUM / LOW on the headline claim, with reasoning.
Then across all papers: where they agree, where they conflict, and what a
sceptical reader would attack first.