tco-search
Skill0 starsChanged 27 days agoMITView on GitHub
Help choose what to buy by comparing total cost of ownership across Amazon and any installed marketplace sources. Use for shopping research, product shortlists, and upfront versus long-term cost comparisons.
In this repository24 other files it gives its agents.
- codebase-designSkill
- improve-codebase-architectureSkill
- uv-pythonSkill
- bw-loginSkill
- project-iconSkill
- recover-mouse-inputSkill
- wizardSkill
- dispatching-parallel-agentsSkill
- handoffSkill
- hill-climbSkill
- opcSkill
- opnSkill
- personal-skillsSkill
- reassign-pmSkill
- repo-lessonsSkill
- trialSkill
- unstuckSkill
- woSkill
- sampleSkill
- find-docsSkill
- liteparseSkill
- browse-xSkill
- frontend-designSkill
- playwrightSkill
The file
We split this skill into the rules it gives an agent
8 rules. About 0 are common, 6 are distinctive, and 2 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈6
Distinctive · estimate
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 12
Read and follow any installed question-style skill (for example wo).
- Very specific: line 13
Use what the user already supplied.
- Distinctive: line 14
Ask at most two rounds, then search with explicit assumptions for unanswered details.
- Distinctive: line 18
Search the relevant Amazon storefront with the browser tools available to the running agent, using ordinary public browsing.
- Distinctive: line 20
Use one search query per source per refined interview, with no retry or query-variation loops.
- Distinctive: line 21
Treat a challenge, block, error, or unavailable tool as no data from that source, never zero results.
- Distinctive: line 26
Use the same currency, keep period, and usage assumptions for every candidate:
- Distinctive: line 34
Use ranges when uncertainty could change the ranking.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
TCO Search
Find the best fit over the user's keep period, not just the lowest sticker price. Research only: never buy, add to cart, or sign in.
Interview first
- Read and follow any installed question-style skill (for example
wo). Otherwise use one numbered list, one line per question, a recommended default per line, and short-answer choices. - Use what the user already supplied. Ask only what changes the search: budget (upfront or lifetime), must-haves and deal-breakers, new/used/refurbished, keep period, running-cost importance and relevant usage, and shipping/tax location. Country and postal code or region suffice; never request a full address.
- Ask at most two rounds, then search with explicit assumptions for unanswered details. If everything is already known, go straight to search.
Search
- Search the relevant Amazon storefront with the browser tools available to the running agent, using ordinary public browsing. No login, CAPTCHA solving, or access-control workarounds. If browser tools are unavailable or Amazon is blocked, say so; do not invent listings.
- Discover and read any installed skills that search other marketplaces, then follow them for additional sources. Do not require a particular marketplace skill or assume it is installed.
- Use one search query per source per refined interview, with no retry or query-variation loops. Opening returned product pages to verify details is fine. A narrower second pass starts only after the user refines the brief.
- Treat a challenge, block, error, or unavailable tool as no data from that source, never zero results. Report the limitation and continue with accessible sources. Say zero results only after a successful search actually returns none.
- Compare exact variants and condition against the must-haves. Record public price, currency, shipping, availability, seller, warranty and return terms, links, and observation date. Mark anything unverified or destination-dependent; never assume free shipping or membership benefits.
Estimate ownership cost
Use the same currency, keep period, and usage assumptions for every candidate:
TCO = price + shipping + estimated sales tax + expected warranty/return/repair cost + running costs over the keep period - expected resale value at the end.
- Upfront means price + shipping + estimated sales tax. Avoid double-counting tax already included in the price; disclose any currency conversion and its rate/date.
- Show the assumptions and label estimates. For expected warranty/return/repair costs, explain likelihood × out-of-pocket cost, accounting for coverage and return fees without counting the same loss twice.
- Include relevant power, consumables, and subscriptions over the keep period. State usage and unit rates; distinguish not applicable from unknown. Missing costs are not automatically zero.
- Estimate end-of-period resale conservatively for the expected age and condition, net of selling costs. Cite evidence and explicitly say when the estimate uses asking prices rather than sold prices. Do not present today's used price as guaranteed future resale.
- Use ranges when uncertainty could change the ranking. Without enough evidence for a cost, mark TCO incomplete rather than fabricate precision. Explain close calls and rank only candidates that meet the requirements; if none can be verified, say so.
Deliver
Give a short ranked table, with assumptions just below it if needed:
| Item | Source | Upfront | TCO | Key assumptions | Link |
State the keep period and currency, disclose source limitations, then give a one-line pick with why it wins for this user. If evidence is insufficient, make the pick conditional or say no defensible pick yet. Offer a narrower second pass, without starting another search automatically.
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.

