Passed 18 of 22 test cases with the skill, and 8 of 22 without it.
Without skill36%
With skill82%
Checks it passed that the model alone missed
The response explicitly informs the user that Chinese databases lack free public APIs.
The response proposes cross-domain method transfer or broadening the method axis as a recommended recourse.
The response specifies recording the user's confirmation as a scope decision bound to their exact statement.
and 10 more in the full report
10 fixed8 passed either way4 still failing0 broke
Pass-rate lift
+45 pp
Tokens
+288%
Agent turns
0%
Response time
-6%
Model gemini-3.6-flashMeasured 8/20/2026v118 of 22 cases passed with skill
Tokens, turns and time describe resource use, not a dollar price.
See it in action
One of 10 test cases the skill fixed. Same model, same prompt, with and without the skill.
1 of 10
Prompt
During automated domain mapping, the retrieved paper set contains exclusively recent papers from the last 3 years, with zero foundational papers older than 8 years. What advisory signal should be generated in the findings output?
What was checked, without and with the skill
Checked
Without skill
With skill
The response specifies emitting a non-critical distribution skew advisory indicating recent heavy coverage missing historical roots.
Missed
Passed
The response suggests performing backward snowballing on top seed papers to recover foundational classic literature.
Passed
Passed
case-22 · gemini-3.6-flash · 8/20/2026 · answers are capped and redacted
What's inside
The instructions your model reads when the skill loads.