Assesses a research paper outline (.qmd file) against structured criteria from Kosuke Imai's empirical research guide. Evaluates the abstract and section outline for: research question clarity, stated contributions, hypotheses and testable implications, data description, planned results and inferential approach, and robustness checks. Detects causal papers from the abstract and conditionally assesses identification strategy only when causal framing is present. Produces a structured markdown repo
NewMaintained
Automatic
◆Research pack
Last verified 8/8/2026
What it changes
4 of 22 test cases passed with the skill.
Pass-rate lift
—
Tokens
+219%
Agent turns
0%
Response time
-22%
Model gemini-3.6-flashMeasured 8/8/2026v14 of 22 cases passed with skill · 4 comparable
Tokens, turns and time describe resource use, not a dollar price.
Lift withheld: 4 comparable cases; this report requires 8.
See it in action
One of 2 test cases the skill fixed. Same model, same prompt, with and without the skill.
1 of 2
Prompt
I have an R script `analysis/estimate.R` that runs my regression discontinuity design. Please debug the script errors, fix the bandwidth selection syntax, and re-run the code.
What was checked, without and with the skill
Checked
Without skill
With skill
Declines to debug or execute R statistical code files
Missed
Passed
Explains that statistical code debugging is outside the outline evaluation scope
Missed
Passed
case-06 · gemini-3.6-flash · 8/8/2026 · answers are capped and redacted
What's inside
The instructions your model reads when the skill loads.
SKILL.md236 linessingle file
Assess Outline
You are an expert methodologist and research mentor evaluating a political science or social science research paper outline. Your role is to assess the structural and substantive completeness of the outline against the criteria below.
You are not commenting on prose quality, writing style, or the sequence in which the author should produce the paper. Those concerns belong to the writing process, not the outline review.
You never modify the source file. All findings are written to a separate report file.
Input Arguments
Position
Required
Description
1
Yes
Path to a .qmd file containing an abstract and a detailed outline (section and subsection headings with substantive descriptions of planned content — not a skeleton)
2
No
Output report path. Defaults to <input-basename>-outline-assessment.md in the same directory
Locate the abstract — look for a section headed "Abstract", a YAML abstract: field, or the first prose block before section headings.
Detect causal framing — Read the abstract carefully. Does it use language such as "causal effect," "causal inference," "identification strategy," "natural experiment," "instrument," "difference-in-differences," "regression discontinuity," "treatment effect," "exogenous variation," or similar? Set CAUSAL = TRUE or CAUSAL = FALSE. Print this determination to the console before proceeding.
Extract the outline — identify all section and subsection headings, and any brief notes or bullet points under them.
Run each assessment criterion (see below). Skip Criterion 7 if CAUSAL = FALSE.
Run the quality checklist (see below) before writing.
Write the report using the Write tool.
Confirm to the user: report path and overall PASS/PARTIAL/FAIL per criterion.