Use when responding to a Comparative Political Studies (CPS) revise-and-resubmit decision — structuring the response letter to multiple anonymous reviewers and the editor. Plans the revision and response; it does not fabricate results or relitigate the decision.
SkillScore86Maintained
Automatic
◆Research pack
Last verified 9/1/2026
What it changes
Passed 19 of 22 test cases with the skill, and 11 of 22 without it.
Without skill50%
With skill86%
Checks it passed that the model alone missed
Uses the tag 【Then】 for the next submission step.
Instructs updating the data availability statement.
Identifies SAGE Track as the resubmission platform.
and 10 more in the full report
8 fixed11 passed either way3 still failing0 broke
Pass-rate lift
+36 pp
Tokens
+35%
Agent turns
0%
Response time
-3%
Model gemini-3.6-flashMeasured 9/1/2026v119 of 22 cases passed with skill
Tokens, turns and time describe resource use, not a dollar price.
See it in action
One of 8 test cases the skill fixed. Same model, same prompt, with and without the skill.
1 of 8
Prompt
We have finalized our structured response document and all manuscript edits for our CPS submission. What final submission platform step is indicated once the revision response preflight is complete?
What was checked, without and with the skill
Checked
Without skill
With skill
Identifies SAGE Track as the resubmission platform.
Missed
Passed
case-22 · gemini-3.6-flash · 9/1/2026 · answers are capped and redacted
What's inside
The instructions your model reads when the skill loads.
SKILL.md93 linessingle file
Rebuttal & R&R Response (cps-rebuttal)
An R&R is a contract: address every point from each reviewer and the editor, show what changed, and
point to where. CPS reviewers are comparativists; they care most about comparative leverage, the rival
explanations, measurement across cases, and inference. Treat the letter as a navigable document the
editor can verify against the revised manuscript.
When to trigger
A decision letter arrives with reviewer reports (R&R, or major/minor revisions)
Planning the revision and the point-by-point response
Reviewers disagree with each other and you must reconcile their demands
A reviewer asks for analyses that would change the paper's identification or scope
Response-letter structure
Opening synthesis. Thank the editor/reviewers; state the 3–4 biggest changes in a short paragraph
so the editor sees the shape of the revision immediately.
Point-by-point. Reproduce each comment verbatim, then Response (what you did) and Location
(section/table/page). Number them (R1.1, R1.2, …) for easy cross-reference.
Show, don't assert. Quote new text, name the new exhibit, give the new estimate — don't just say
"we have addressed this."
Handle disagreement. Where reviewers conflict, state the trade-off and your principled choice;
ask the editor to adjudicate if needed. Where you decline a change, give a reasoned, evidence-based
defense — politely, without relitigating the decision.
CPS-specific revision priorities
Comparative leverage / travel. If a reviewer doubts the comparison or generalization, strengthen
the design argument or the scope conditions (cps-research-design, cps-theory-building) — do not
just add caveats.
Rival explanations. New robustness or case evidence that adjudicates the rival the reviewer raised;
state the adjudication sentence explicitly.
Measurement across cases. If asked, show robustness to alternative codings (V-Dem/Polity/etc.) and
defend comparability.
Inference. Add few-cluster corrections, randomization inference, or multiple-testing adjustments if
challenged; report honest CIs.
Transparency. Update the replication package and data availability statement to match every new
result (cps-transparency-and-data).
Use when writing the response to an American Journal of Political Science (AJPS) revise-and-resubmit. The revised manuscript stays double-blind, the response memorandum is submitted anonymized, and any new analyses must keep the replication package re-runnable for the third-party verifier at acceptance. Structures the response; it does not fabricate new results.