Find agent skills by outcome

- For each option: WHY it exists

HOW it works

- Share recommendation with reasoning

then ask user to decide

- When multiple valid approaches: speculate

think aloud

- If hypotheses disproven

generate new ones from gathered information

- Policies

rules

2. Apply Deepthink Protocol (reason about dependencies

risks

- Policy rules

mandatory prerequisites

- Identify gaps

ambiguities

If you can track internally

do it internally.

If not

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

- Scan package.json

Cargo.toml

- Check for .claude/ folder (agents

commands

- Look for CLAUDE.md

AGENTS.md

Turkish Cats hanging out nearby of Galata Tower

Turkish Cats hanging out nearby of Galata Tower, vertical

Ultrathinker

# Ultrathinker

Before any action

gather context and enhance the request internally:

- Use soft

warm tones to evoke a familial and welcoming atmosphere.

- Use a background setting that complements the family theme

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

- Blend the features

expressions

- Provide access to a variety of games

from retro to modern.

3. Deliver results in Markdown

JSON/YAML

2. Propose enhancements to tools

processes

1. Provide unit

integration

- Prioritize bugs based on severity

user impact