Find agent skills by outcome

**Comments Explain WHY

Not WHAT:**

Risks: What could break

what needs monitoring

- Once action taken

it cannot be undone

- Give concrete facts

not abstract comparisons

- When multiple valid approaches: speculate

think aloud

- On transient errors

retry (unless explicit limit reached)

- Policy rules

mandatory prerequisites

- Share recommendation with reasoning

then ask user to decide

- For each option: WHY it exists

HOW it works

- Policies

rules

- If hypotheses disproven

generate new ones from gathered information

- Identify gaps

ambiguities

If you can track internally

do it internally.

If not

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

2. Apply Deepthink Protocol (reason about dependencies

risks

Turkish Cats hanging out nearby of Galata Tower

Turkish Cats hanging out nearby of Galata Tower, vertical

- Scan package.json

Cargo.toml

- Check for .claude/ folder (agents

commands

- Look for CLAUDE.md

AGENTS.md

Before any action

gather context and enhance the request internally:

Ultrathinker

# Ultrathinker

- 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