Passed 16 of 22 test cases with the skill, and 3 of 22 without it.
Without skill14%
With skill73%
Checks it passed that the model alone missed
requires peer comparison across key financial ratios.
mandates a monitoring checklist for thesis validation.
requires an A-E source quality rating scale for citations.
and 17 more in the full report
15 fixed1 passed either way4 still failing2 broke
Pass-rate lift
+59 pp
Tokens
+120%
Agent turns
0%
Response time
-1%
Model gemini-3.6-flashMeasured 8/23/2026v116 of 22 cases passed with skill
Tokens, turns and time describe resource use, not a dollar price.
See it in action
One of 15 test cases the skill fixed. Same model, same prompt, with and without the skill.
1 of 15
Prompt
I'm planning to analyze Kweichow Moutai (600519) for potential inclusion in my portfolio. Please ask me for the necessary background context, holding horizon, and preferred analytical depth first, waiting for my response before drafting the complete research outline.
What was checked, without and with the skill
Checked
Without skill
With skill
asks diagnostic questions across required investment parameter categories.
Passed
Passed
includes an explicit disclaimer stating that this is not financial or investment advice.
Missed
Passed
case-02 · gemini-3.6-flash · 8/23/2026 · answers are capped and redacted
What's inside
The instructions your model reads when the skill loads.
SKILL.md242 linessingle file
Stock Question Refiner
Role
You are a Stock Investment Research Question Refiner specializing in crafting structured investment research prompts. Your primary objectives are:
Ask clarifying questions first to understand the user's investment context, style, and research needs
Generate structured research prompts that follow professional investment due diligence standards
Eliminate vague requests by defining clear research parameters
Core Directives
Do Not Provide Investment Advice: Focus on research planning, not stock recommendations
Be Explicit & Skeptical: If the user's request is unclear, ask targeted questions
Enforce Structure: Ensure every research request has clear scope, constraints, and output requirements