Use when revising an American Economic Review: Insights (AER: Insights) short-format manuscript to lead with the result, fit the hard word cap (7,000 words minus 200 per exhibit), and write the ≤100-word abstract. Enforces the length budget and the lead-with-the-result arc; it does not change the analysis.
SkillScore91Maintained
Automatic
◆Writing & Editing pack
Last verified 8/15/2026
What it changes
Passed 20 of 22 test cases with the skill, and 10 of 22 without it.
Without skill45%
With skill91%
Checks it passed that the model alone missed
Names aeri-replication-package as the next step value.
Prohibits p-value asterisks anywhere in tables or text.
Outputs the summary using block header 【First paragraph】.
and 13 more in the full report
10 fixed10 passed either way2 still failing0 broke
Pass-rate lift
+45 pp
Tokens
+23%
Agent turns
0%
Response time
-36%
Model gemini-3.6-flashMeasured 8/15/2026v120 of 22 cases passed with skill
Tokens, turns and time describe resource use, not a dollar price.
See it in action
One of 10 test cases the skill fixed. Same model, same prompt, with and without the skill.
1 of 10
Prompt
If an author includes the maximum allowed 5 exhibits in an AER: Insights paper, what is the maximum word budget for the main text?
What was checked, without and with the skill
Checked
Without skill
With skill
Calculates the main text ceiling at 5 exhibits as 6,000 words.
Missed
Passed
case-22 · gemini-3.6-flash · 8/15/2026 · answers are capped and redacted
What's inside
The instructions your model reads when the skill loads.
SKILL.md71 linessingle file
Writing Style — Lead With the Result, Fit the Cap (aeri-writing-style)
When to trigger
The intro buries the headline result under setup, literature, or methods
The draft is over the word cap and needs cutting, not just polishing
The abstract exceeds 100 words or does not state the result
Late-stage polish once identification, robustness, and exhibits have settled
The two hard constraints (检索于 2026-06；以官网为准)
Word cap: main text ≤7,000 words with no exhibits; each exhibit subtracts 200 words; ≤5 exhibits (so ≥6,000 at the cap). The count includes body, footnotes, endnotes, and in-paper appendices; it excludes title, authors, abstract, acknowledgement footnote, references, exhibits, and the Supplemental Appendix. Over the cap = returned unreviewed.
Abstract: ≤100 words. It must state the result, not promise it.
Treat the cap as a design constraint from the first draft, not a final trim. Every sentence competes with the insight for space.
The AER: Insights arc (lead with the result)
Unlike a long paper that builds to its result, an AER: Insights paper states the answer in the first paragraph:
Sentence 1–2: the question and the headline result with its magnitude and uncertainty (estimate + SE / confidence set), no asterisks.
Sentence 3–4: why it matters / the prior it overturns (the surprise).
One short paragraph: how it is identified (the design or the model's key restriction), in plain terms.
One sentence: the single most important robustness/scope caveat, with a pointer to the appendix.
Brief roadmap or none — the paper is short enough that heavy signposting wastes words.
The body then delivers exactly that arc: setup (tight), identification, result, the one or two checks, a short discussion. No second result, no survey, no extended derivations in-text.
Cutting to the cap (in priority order)
Cut secondary results to the Supplemental Appendix (aeri-robustness).
Compress the literature to one or two paragraphs (aeri-literature-positioning).
Move proofs/derivations and estimation detail to the appendix (aeri-theory-model).
Delete throat-clearing ("It is well known that…", over-signposted roadmaps).
Merge exhibits to reclaim 200 words each (aeri-tables-figures).
Tighten prose last — shorter sentences, fewer hedges, active voice.
Writing the ≤100-word abstract
State the result with its magnitude, the setting, and the identification in one breath.
Use when building or trimming the exhibits of an American Economic Review: Insights (AER: Insights) short-format manuscript to the hard five-exhibit budget, one page each, no significance asterisks. Designs the minimum exhibit set; it does not run the analysis behind them.