Find agent skills by outcome

- Evaluate the structure and data of website databases

identifying trends or anomalies.

Risks: What could break

what needs monitoring

Smells: Tech debt

workarounds

- Give concrete facts

not abstract comparisons

- Share recommendation with reasoning

then ask user to decide

- For each option: WHY it exists

HOW it works

How: Key implementation decisions (patterns used

structure chosen)

- Once action taken

it cannot be undone

- Policies

rules

- When multiple valid approaches: speculate

think aloud

- On transient errors

retry (unless explicit limit reached)

- If hypotheses disproven

generate new ones from gathered information

- Check for .claude/ folder (agents

commands

- Look for CLAUDE.md

AGENTS.md

- Scan package.json

Cargo.toml

If not

create `todos.txt` at project root—update as you go

If you can track internally

do it internally.

- Identify gaps

ambiguities

- Policy rules

mandatory prerequisites

2. Apply Deepthink Protocol (reason about dependencies

risks

- Use a background setting that complements the family theme

such as a cozy living room or an outdoor garden scene.

Turkish Cats hanging out nearby of Galata Tower

Turkish Cats hanging out nearby of Galata Tower, vertical

Before any action

gather context and enhance the request internally:

Ultrathinker

# Ultrathinker