Find agent skills by outcome

- Scikit-learn (or light ML

no heavy black-box models initially)

- Damage status (None

Minor

At minimum

support the following inputs:

- Prefer robust estimators (median

trimmed mean

- Fuel type (Petrol

Diesel

You may propose alternatives

but default to:

Data

- Is transparent and data-driven

not speculative

- Manipulation

emotional pricing

- Turkey-specific market dynamics (inflation

taxes

You will work in an agent-style

vibe coding approach:

Data

- Users want a fair

data-driven estimate of their car’s real market value.

- Existing marketplaces (e.g.

classified ad platforms) have highly volatile

You are building a full-stack web and mobile application inspired by the Kelley Blue Book – What's My Car Worth? concept

but strictly tailored for the Turkish automotive market.

Data

- Develop the backend using ${backendTechnology:Node.js}

ensuring secure and efficient data handling.

Education

- Evaluation: Explain how the workshop will be evaluated

mentioning options like journals

- Resources: Identify and prepare texts

videos

- Execution: Detail how the workshop will be conducted (online

virtual

- Develop a clear plot structure with a beginning

middle

implementation_note: Typography scale

alignment

noise: Very low

studio clean

imperfections: Very subtle

realistic — not overly polished CGI

fidelity: High material accuracy (paint

carbon fiber

tone: High-performance

premium motorsport

emotion: Precision

speed