Use when building the theoretical argument and hypotheses of a Communication Research (CR) manuscript into a testable, mechanism-level contribution. CR rewards explicit constructs, a stated communication process, and numbered hypotheses with mediation/moderation, not a bare effect. Structures the argument; it does not run analyses.
SkillScore96Maintained
Automatic
◆Research pack
Last verified 9/1/2026
What it changes
Passed 24 of 25 test cases with the skill, and 9 of 25 without it.
Without skill36%
With skill96%
Checks it passed that the model alone missed
Output uses the bracketed tag 【Core claim】.
Output includes the bracketed tag 【Mechanism】.
Output includes the bracketed tag 【Core claim】.
and 17 more in the full report
16 fixed8 passed either way0 still failing1 broke
Pass-rate lift
+60 pp
Tokens
+26%
Agent turns
0%
Response time
-13%
Model gemini-3.6-flashMeasured 9/1/2026v124 of 25 cases passed with skill
Tokens, turns and time describe resource use, not a dollar price.
See it in action
One of 16 test cases the skill fixed. Same model, same prompt, with and without the skill.
1 of 16
Prompt
I am preparing a communication manuscript on how customized privacy warnings in mobile apps influence user compliance. R1 noted that my study shows a direct relationship but lacks a tested theoretical process. Could you structure my conceptual framework? Please lay out a single-sentence primary claim, the explanatory communication mediator, precise definitions of the key constructs (distinguishing them from closely related terms), numbered hypotheses (covering direct effect, mediation path, and moderation), the relevant audience and context boundaries, a warrant explaining how the study design justifies the causal sequence, and the recommended next research step.
What was checked, without and with the skill
Checked
Without skill
With skill
Output includes the bracketed tag 【Core claim】.
Missed
Passed
Output includes the bracketed tag 【Mechanism】.
Missed
Passed
Output specifies commres-research-design under the 【Next】 section.
Missed
Passed
case-01 · 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.md91 linessingle file
Theory & Hypotheses (commres-theory-building)
At CR a result is not a contribution until it is attached to a testable communication mechanism.
This skill turns an idea into theory: precise constructs, an explicit process, and numbered
hypotheses that the design will test. CR is a hypothetico-deductive journal — a paper that documents
an effect without a tested mechanism rarely clears the bar.
When to trigger
The empirics are strong but the "so what / why" is thin
A reviewer said the paper is "atheoretical," "ad hoc," or "just a main effect"
You need to state constructs, mechanisms, and predictions explicitly before collecting data
Connecting your work to (or extending) an established communication theory
Build the argument (theory → mechanism → hypotheses)
Concept. Define the key constructs precisely (e.g., exposure, framing, presence, perceived
norms); distinguish each from its near neighbors so measurement is unambiguous.
Mechanism. State the communication process: who sends/receives what, through which channel,
with what cognitive/affective/social step that produces the outcome. This is the mediator.
Boundary conditions. Name the moderators — which audiences, messages, and contexts strengthen
or reverse the effect. CR rewards "for whom / when," not just "whether."
Hypotheses. Translate the mechanism into explicit, numbered hypotheses (and research
questions where theory is thin): H1 (effect), H2 (mediation path), H3 (moderation). These become
the tests in commres-research-design and commres-data-analysis.
The mediation/moderation core (CR-specific)
CR is the home of process models — message → mediator → outcome, conditioned by moderators. State
the causal ordering the model assumes and how the design licenses it: a mediator measured on the
same cross-sectional wave as the outcome cannot carry a causal-process claim (defer the temporal
warrant to design). Pre-specify the indirect-effect test; do not fit a PROCESS model post hoc and
narrate it as theory.
The theory-contribution bar at CR (calibration anchor, hedged)
A common substantive rejection is "theory cited but not advanced." Citing a theory is not moving
it. Calibrate where a paper sits:
| Level | What the paper does to theory | CR verdict (typical) |