Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →brycewang-stanford/aeri-literature-positioning
Use when positioning the contribution of an American Economic Review: Insights (AER: Insights) short-format manuscript against the closest published work in one or two tight paragraphs. Stakes the gap concisely; it does not write the full literature review (there is no room for one).
Passed 21 of 22 test cases with the skill, and 3 of 22 without it.
and 38 more in the full report
19 fixed2 passed either way0 still failing1 broke
Tokens, turns and time describe resource use, not a dollar price.
One of 19 test cases the skill fixed. Same model, same prompt, with and without the skill.
Our AER: Insights manuscript on CEO pay isolates shareholder monitoring from manager entrenchment, which previous studies bundled together. Generate the positioning output block.
| Checked | Without skill | With skill |
|---|---|---|
| States the contribution as isolating or separating mechanisms that were previously bundled. | Passed | Passed |
| Limits the closest prior work block to at most 2 papers. | Missed | Passed |
The instructions your model reads when the skill loads.
AER: Insights gives you no room for a survey. Positioning must be done in one or two paragraphs that do exactly one job: pin the single closest prior result and state precisely what your one insight adds, corrects, or overturns. Because the whole paper is one idea, the literature task is narrow — locate that idea on the frontier, not map the whole field. The word budget is shared with everything else (≤7,000 words minus exhibits), so every citation must earn its place.
aeri-topic-selection.| Gap type | One-line framing | Watch-out |
|---|---|---|
| New result on an open question | "It was unknown whether …; we show …" | be sure it is genuinely open, not just unaddressed-by-you |
| Overturning a believed result | "Prior work concluded X; with [better design/data] we find not-X" | engage the prior paper's design fairly |
| Sharpening / separating channels | "Prior work left A and B bundled together; we isolate ..." | show the separation is the contribution |
| Method that changes the answer | "Applying [method] to [setting] reverses the standard conclusion" | the payoff, not the method, is the contribution |
In a long paper the literature has its own section; in AER: Insights it should live inside the
introduction — typically the second and third sentences after the headline result
(aeri-writing-style). The reader meets the result, then immediately
learns what was believed before and what your one insight changes. A standalone "Related Literature"
section is a smell that the paper is too long or trying to be an AER paper.
Because AER: Insights is general-interest, the closest paper you confront should be one a broad
test cases passed, vs 3 / 22 without
Published skills whose descriptions closely match this one, with their headline benchmark lift.
More skills from the same source and nearby in the registry, with their headline benchmark lift.