Use when designing or auditing the experimental plan for a Cancer Cell (Cell Press) study — orthogonal validation across in vitro, in vivo, and human tumor systems, with controls, replicates, power, randomization, and blinding. It plans design; it does not run statistics or write figures.
SkillScore77Maintained
Automatic
◆Research pack
Last verified 8/31/2026
What it changes
Passed 17 of 22 test cases with the skill, and 16 of 22 without it.
Without skill73%
With skill77%
Checks it passed that the model alone missed
Uses bracketed tag formatting for section headers.
Routes the next step field to cc-reporting-standards.
Requires a non-targeting scramble control matched to delivery.
and 1 more in the full report
3 fixed14 passed either way3 still failing2 broke
Pass-rate lift
+5 pp
Tokens
+21%
Agent turns
0%
Response time
-14%
Model gemini-3.6-flashMeasured 8/31/2026v117 of 22 cases passed with skill
Tokens, turns and time describe resource use, not a dollar price.
See it in action
One of 3 test cases the skill fixed. Same model, same prompt, with and without the skill.
1 of 3
Prompt
I am outlining a study on epigenetic silencing in glioblastoma utilizing shRNA knockdown in cell lines and patient-derived organoids. I want to make sure the experimental rigor meets top journal standards. Please analyze my research plan and format your response with headings for: Causal claim layer, Orthogonal systems planned, Controls per perturbation, Replicates, Animal rigor, Human anchor, Gaps to close before submission, and Next step.
What was checked, without and with the skill
Checked
Without skill
With skill
Uses bracketed tag formatting for section headers.
Missed
Passed
Recommends using at least two independent shRNA sequences.
Passed
Passed
Routes the next step field to cc-reporting-standards.
Missed
Passed
case-02 · gemini-3.6-flash · 8/31/2026 · answers are capped and redacted
What's inside
The instructions your model reads when the skill loads.
SKILL.md81 linessingle file
Experimental Design (cc-study-design)
When to trigger
The mechanism rests on a single system (e.g., cell lines only)
Reviewers will ask "is this true in vivo?" or "does it hold in patients?"
Controls, replicates, or sample-size logic are unclear
Planning mouse / PDX / organoid experiments and unsure about rigor elements
Orthogonal validation ladder
Cancer Cell expects a mechanism corroborated across independent, complementary systems. Build the strongest ladder the biology allows:
Patient tumor samples, TMAs, scRNA-seq, public cohorts (TCGA), outcome data
Translational anchor / clinical relevance
A Cancer Cell paper typically spans in vitro + in vivo + a human anchor. Decide early which layers carry the causal claim and which provide corroboration.
Controls (the parts reviewers attack)
Genetic perturbation: include rescue / add-back; use ≥2 independent sgRNAs or shRNAs (or rule out off-target); non-targeting / scramble control matched to delivery.
Pharmacology: vehicle control, dose-response, on-target genetic phenocopy, and where possible a resistant mutant or analog.
In vivo: isotype/vehicle controls, littermate or co-housed controls for GEMMs, tumor-volume and endpoint pre-defined.
Antibody / staining: isotype, KO/KD-validated, single-stain compensation for flow.
Replicates: biological vs technical
Biological replicates = independent biological units (separate mice, independent cell passages/cultures, distinct patients). These define n.
Technical replicates (duplicate wells, repeat measurements) reduce measurement noise but do not increase n.
Report both clearly; never inflate n with technical replicates (see cc-statistics).