product-research
Skill1 starsChanged 2 months agoMITView on GitHub
Research products to help the user decide what to buy. Use when the user wants recommendations, comparisons, or help choosing between products, for personal or work purchases, at any budget. Trigger on requests like 'what should I buy', 'help me find a', 'recommend a', 'which is better', or any purchase decision, even if no budget or use case has been given yet.
14 other projects use a version of this.
Adopted by
Most recently taken up by
- darkweb19/skillstheir own version, first seen Nov 8, 2026
- ethanratnofsky/skillstheir own version, first seen Nov 8, 2026
- newkirkjosh/skillstheir own version, first seen Oct 25, 2026
- liangdabiao/amazon-sorftime-research-MCP-skilltheir own version, first seen Oct 13, 2026
- alexander-zuev/agent-skillstheir own version, first seen Oct 13, 2026
- rolandtolnay/llm-toolkittheir own version, first seen Oct 11, 2026
Other versions14 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 14
- liangdabiao/amazon-sorftime-research-MCP-skill★ 1.1ksame name
- AlpacaLabsLLC/skills-for-architects★ 420same name
- mlopscommunity/Coding-Agents-Conference-skills★ 38same name
- Hassaan146/claude-skills★ 18same name
- Ethan-paseya/amazon-sellersprite-research-MCP-skill★ 2same name
- Hussain-Elfalah/Agentic-Skills★ 2same name
- 199-biotechnologies/product-research★ 1same name
- mahiunfiltered/claude-code-skills★ 1same name
- newkirkjosh/skills★ 1same name
- rolandtolnay/llm-toolkit★ 1same name
4 more not listed
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository10 other files it gives its agents.
- notes/agents.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- claude-project-structureSkill
- code-philosophySkill
- writing-documentationSkill
- fact-checkingSkill
- professional-communicationSkill
- skill-creationSkill
- skill-refactoringSkill
- tooling-researchSkill
The file
We split this skill into the rules it gives an agent
3 rules. 0 are common, 3 are distinctive, and 0 are very specific to this project.
0
Common
Most of the 13 edited versions we read keep these.
3
Distinctive
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 8
Never rely on training data alone for prices, specs, or "best in class" picks.
- Distinctive: line 9
If budget, use case, or a hard requirement is missing, ask the user rather than assuming it.
- Distinctive: line 46
Every current guide agrees on one thing regardless of brand: prioritize RAM over CPU speed or screen resolution for coding workloads, since …
“Common” is measured: kept by at least 6 in 10 of the 13 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Rules
- Never rely on training data alone for prices, specs, or "best in class" picks. These change constantly, so verify with a search performed in this conversation before stating them as current.
- If budget, use case, or a hard requirement is missing, ask the user rather than assuming it.
Steps
- Confirm what's actually unclear before searching. Do not assume budget, use case, or personal-vs-work if the prompt doesn't say. Use the ask_user_input_v0 tool to confirm the missing pieces in one short round, for example: intended use (personal / work / gift), budget tier (low / mid / high / no limit), and any hard requirements (brand, size, compatibility, must-haves). If the prompt already answers one of these, don't ask again, just confirm the others. Skip the tool entirely if the user already gave enough detail to work with (specific budget, clear use case, named requirements) - asking anyway wastes their time.
- Make zero assumptions about the market. Know what in your training is stale. Product lineups, prices, and "best in class" picks change constantly, so treat any specific model, price, or spec claim as something to verify now, not recall.
- Search the web. Look at current buying guides, retailer listings, and expert reviews. Favor recent, independent, specific sources (reputable review sites, hands-on comparisons) over generic listicles or SEO-driven affiliate content. Cross-check prices and specs across at least two sources since prices shift and listings go stale.
- Narrow to a short list. Don't dump every option that exists. Pick 2-4 real candidates that fit the confirmed budget and use case, and say why each one made the cut over the alternatives.
- Report directly. For each candidate: what it is, current price, why it fits the stated use case and budget, and the one or two tradeoffs against the others. Be direct and brief, skip filler like "this is a great choice for anyone."
- State assumptions and gaps at the end. If anything wasn't confirmed (e.g. the user skipped a question, or a spec couldn't be verified), say so plainly rather than quietly filling it in.
Example
User: "I need a new laptop for work."
Claude: [Uses ask_user_input_v0 to confirm: primary work tasks (e.g. general office / creative or video work / heavy coding or dev work), budget tier (under $800 / $800-1500 / $1500+), OS preference if any]
User selects: heavy coding/dev work, $800-1500, no OS preference.
Claude: [Searches current buying guides and reviews for dev-focused laptops in that price range, cross-checks recommendations and specs across multiple sources]
For dev work in the $800-1500 range, current reviews consistently point to three real options:
Lenovo ThinkPad T14 / T16 - widely recommended across programming-laptop guides right now for its durable keyboard, mostly non-soldered components (so RAM/storage are upgradeable later), and it runs Linux well. Tradeoff: bulkier and heavier than a MacBook, and the base config often needs a RAM bump to hit 32GB.
Framework Laptop 13 - built to order, and the standout pick if repairability matters: swap ports, RAM, storage, or the mainboard yourself later instead of replacing the whole machine. Tradeoff: shorter reliability track record than Lenovo, and pricing depends heavily on the configuration chosen.
Apple MacBook Pro (M-series) - rated as the top overall pick in more than one current comparison for raw performance and display quality, if macOS is acceptable. Tradeoff: sits at the top of this budget range and RAM can't be upgraded after purchase, so buy more than you think you need up front.
Every current guide agrees on one thing regardless of brand: prioritize RAM over CPU speed or screen resolution for coding workloads, since IDE plus browser plus containers plus AI tooling running at once is what actually strains a machine day to day.
What I didn't confirm: exact current prices and configurations change often enough that I'd verify the specific model and RAM configuration against the retailer at the time of purchase, and whether discrete graphics matters (e.g. for local ML work) - let me know and I can narrow further.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

