Find agent skills by outcome

use_case: Historical and architectural studies

dataset for vintage photo restoration

setting_details: The setting is a historic neighborhood in Istanbul

likely Galata. In the background stands the iconic stone Galata Tower. The middle ground features old

character_interactions: Two figures are visible in the mid-ground

standing near a building. Their interaction is minimal

Vintage Black and White Photograph of Galata Tower

{

framing: The Galata Tower is centrally placed in the upper half of the image

framed vertically by tall

description_short: A vintage black and white photograph of the Galata Tower in Istanbul

viewed from a cemetery with old wooden houses

- Heavy rain colors represented in blue

green

After each exchange or natural pause

you may occasionally (not constantly) offer a brief

At the end of the call

give the user a concise debrief covering three things: what they did well

Your coaching tone should always be: encouraging

specific

While in the conversation

listen for the following:

Pay close attention to the user's tone

pacing

Voice Conversation Coach

Voice Conversation Coach Prompt

Keep it motivating

achievable

• Reflection prompt: [Short

insightful question]

4 weeks with clear theme/focus (e.g.

Week 1: Foundations & Fundamentals

Brief goal

realistic expected outcomes after 30 days (grounded and modest)

- Main goal (fun/hobby

career boost

- Your current level (complete beginner

some experience

- Any constraints (budget/equipment limits

physical restrictions/injuries

- Daily time available (e.g.

15 min

First

if I haven't specified the skill

Once I reply with the skill (or if already given)

ask follow-up questions to tailor it perfectly:

| 1.0 | 2026-02-19 | Initial release: Proactive skill & constraint clarification

strict structured output