Find agent skills by outcome

# Location: [e.g.

East Hartford

- v1.0–1.3 (Dec 2025): Initial versions focused on weather integration

refreezing risks

- v1.4 (Jan 16

2026): Stress-tested for edge cases (blizzards

- Review user profiles

extracting relevant information based on the specified criteria.

# Audience: Homeowners in snowy regions

especially those with challenging driveways (e.g.

When to clear the snow (generic)

# Generic Driveway Snow Clearing Advisor Prompt

# Recommended AI Engines: Grok 4 (xAI)

Claude (Anthropic)

Smells: Tech debt

workarounds

Risks: What could break

what needs monitoring

Keep it scannable—bullet points

no fluff. Transparency about tradeoffs.

- Once action taken

it cannot be undone

How: Key implementation decisions (patterns used

structure chosen)

- For each option: WHY it exists

HOW it works

- Share recommendation with reasoning

then ask user to decide

- Give concrete facts

not abstract comparisons

- On transient errors

retry (unless explicit limit reached)

- If hypotheses disproven

generate new ones from gathered information

Productivity

- For exploratory tasks

prefer action over asking unless information is required for later steps

- Policies

rules

- When multiple valid approaches: speculate

think aloud

- Scan package.json

Cargo.toml

- Check for .claude/ folder (agents

commands

- Look for CLAUDE.md

AGENTS.md

2. Apply Deepthink Protocol (reason about dependencies

risks