Find agent skills by outcome

Risks: What could break

what needs monitoring

Audience: Homeowners in snowy regions

especially those with challenging driveways (e.g.

- If hypotheses disproven

generate new ones from gathered information

- For each option: WHY it exists

HOW it works

- Share recommendation with reasoning

then ask user to decide

- On transient errors

retry (unless explicit limit reached)

- Once action taken

it cannot be undone

- When multiple valid approaches: speculate

think aloud

How: Key implementation decisions (patterns used

structure chosen)

- Give concrete facts

not abstract comparisons

- Policies

rules

Ultrathinker

If you can track internally

do it internally.

- Identify gaps

ambiguities

- Check for .claude/ folder (agents

commands

PromptBeginner5 minmarkdownQuality: 24

Turkish Cats hanging out nearby of Galata Tower

Turkish Cats hanging out nearby of Galata Tower, vertical

If not

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

Before any action

gather context and enhance the request internally:

- Look for CLAUDE.md

AGENTS.md

- Policy rules

mandatory prerequisites

- Scan package.json

Cargo.toml

2. Apply Deepthink Protocol (reason about dependencies

risks

- Use soft

warm tones to evoke a familial and welcoming atmosphere.