Use when building the exhibits of an Accounting, Organizations and Society (AOS) manuscript — data-inventory and evidence tables for qualitative work, cell-means and process-test tables for experiments, and figures that carry theoretical weight — in Elsevier-compatible format. Builds exhibits; it does not run the analysis (aos-data-analysis).
SkillScore100Maintained
Automatic
◆Research pack
Last verified 8/15/2026
What it changes
Passed 22 of 22 test cases with the skill, and 15 of 22 without it.
Without skill68%
With skill100%
Checks it passed that the model alone missed
States that fixed-effects structures must be documented in table notes
Flags institutional logos in figures as double-anonymized review violations
Requires each exhibit to prove a core theoretical sentence of the manuscript
and 4 more in the full report
7 fixed15 passed either way0 still failing0 broke
Pass-rate lift
+32 pp
Tokens
+16%
Agent turns
0%
Response time
-30%
Model gemini-3.6-flashMeasured 8/15/2026v122 of 22 cases passed with skill
Tokens, turns and time describe resource use, not a dollar price.
See it in action
One of 7 test cases the skill fixed. Same model, same prompt, with and without the skill.
1 of 7
Prompt
I am finalizing a field study on management control systems for submission to AOS. Could you review my manuscript's draft exhibits and figures? Please provide feedback broken down into: a summary listing each exhibit and its specific main purpose, an assessment of the qualitative data tracking tables (like interviews and field observations), a check on whether all titles and footnote definitions make the tables readable on their own, a verification of participant blinding and formatting rules, and the indicated follow-up stage.
What was checked, without and with the skill
Checked
Without skill
With skill
Evaluates whether each listed exhibit supports a single specific theoretical claim
Missed
Passed
Recommends tabulating qualitative data sources by participant role and interview count
Passed
Passed
Requires defining acronyms and variable symbols directly within the table notes
Passed
Passed
case-01 · 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.md68 linessingle file
Tables & Figures (aos-tables-figures)
When to trigger
A field paper has long quotations but no exhibit showing the evidentiary base
Experimental results lack a clean cell-means or contrast table
A figure decorates rather than argues (or a needed process figure is missing)
Formatting must be reconciled with Elsevier submission requirements
Exhibits by tradition — what AOS reviewers look for
Qualitative field / historical papers
Data-inventory table: interviews (number, roles, duration, period), observations (events, hours), documents (types, spans) — the single exhibit that preempts "how much data is behind this?"
Evidence table(s): conceptual claim or second-order theme in one column, representative field excerpts in the other, so the interpretive chain is inspectable without bloating the text.
Coding-structure exhibit where the paper's method warrants it: first-order codes → second-order categories → aggregate concepts, showing movement from informant language to theory.
Timeline or process figure for change studies: phases, key events, and where the accounting practice intervened.
Anonymize consistently — pseudonyms and role labels in exhibits must match the text and betray no site identity.
Experimental / survey papers
Cell-means table: means, SDs, n per condition; the predicted pattern visible at a glance.
Contrast/ANOVA table plus process results (mediation with bootstrap CIs) as exhibits, not prose asides.
An interaction figure with error bars when a moderation is the theory's signature; axis units interpretable, no truncated scales that exaggerate effects.
Survey papers: construct table (items, sources, reliabilities) and a correlation matrix.
Archival-with-theory papers
Descriptives, the main estimation table with clustering and fixed-effects structure stated in the notes, and magnitudes interpreted in institutional terms.
Self-containment and format
Every exhibit readable alone: title stating what it shows, sample/period, definitions, and the meaning of any markers in the notes.
Number tables and figures consecutively; place callouts in the text; Elsevier prefers editable source files for tables and high-resolution figure files at final stages — confirm current artwork specs in the AOS Guide for Authors before submission.
Keep exhibits inside the double-anonymized envelope: no author names in captions, file names, or embedded metadata.
Ration exhibits: each one must carry evidence or structure the argument; move genuinely supplementary material to an appendix or online supplement.
The one-sentence test
For every exhibit, write the sentence it proves ("Table 2 shows the interpretation rests on 74 interviews across four roles over 26 months"; "Figure 3 shows the effect reverses under high accountability"). If no sentence exists, cut the exhibit; if the sentence is central and no exhibit backs it, build one.