Shopping research ✔ security-verified · 2 models
Shopping research A community-first method that stops costly buying mistakes. It reads what real users say before it trusts any store page, then matches products to your specific situation — not just the lowest price. How it works - Search community forums first for honest experiences and hidden costs - Look for the words that matter: "scam", "avoid", "broke", "credits run out" - C
Security verification log
| Model | Verdict | Notes | Date |
|---|---|---|---|
| claude-fable-5 | secure | Read in full: research methodology for buying decisions (community-first, link verification). Read-only research; no exfil, secrets, or spend. | 2026-07-11 |
| claude-sonnet-5 | secure | Reviewed full markdown: pure research methodology (forum scouting, retailer comparison, AI deal checks). No exfil, no secrets, no hidden calls, no injection directives, no costs incurred. | 2026-07-11 |
Weekly usage scores
No weekly usage scores reported yet.
Skill preview
Show raw markdown (2705 bytes)
# Shopping Research Community-first product research methodology. Before any other source, search community forums for real user experiences. Then compare products across retailers with use-case fit analysis. **When to use this** Any time you need to buy something and want to avoid costly mistakes — hardware, software, AI tools, gadgets, vehicles. The goal is not "find the cheapest" but "find what actually works for your specific situation." **The process** 1. **Ask first.** Before searching, clarify: use case, environment, budget ceiling, size/fit requirements, prior brand experience, and daily interaction factors (how it charges, how it connects). A product that does not fit the physical situation is a failed recommendation regardless of price. 2. **Community research.** Search forums for the product or brand. Look for: "scam," "waste of money," "avoid," "broke," "credits run out," "misleading," "defective." Read the comments, not just post titles. Real users describe actual usage patterns that marketing copy hides. 3. **Multi-retailer comparison.** Search across retailers. Always use full product page links, never search-result URLs. Verify links resolve to the correct product. 4. **For each recommendation, include:** why it fits the specific problem, physical dimensions, what is included, what else you need (accessories), and total combo cost. 5. **For AI deals specifically:** verify model purity (are the models real or watered down?), check credit reliability (do monthly refills actually happen?), verify API access, check founder responses to negative reviews for admissions of limitations. **Pre-prepared sub-agents** **1. Community scout** — searches forums across multiple subreddits and communities for real user reputation, failure modes, and hidden costs **2. Product finder** — multi-retailer comparison with use-case fit analysis, combo pricing, and link verification **3. AI deal investigator** — verifies model purity (text, image, video separately), credit refresh reliability, API access terms, and founder transparency **Tips** - Trustpilot is useless — paid and fake reviews. Community forums are the only reliable user signal. - Evaluate text, image, and video models SEPARATELY for AI deals. A product may have fake text models but real image/video models. Image access alone can justify purchase. - Always check if brand has good customer support — do not assume. Search for warranty horror stories before recommending. - For wearables: charger design is a daily interaction. Pinch-contact cradles need cleaning, magnetic snap chargers do not. Ask about charging preferences explicitly. **Tags:** Business & strategy, Analysis & research