nexscope-ai/eCommerce-Skills: 157 Markdown Files That Teach an Agent to Reason About a Store
E-commerce skills for AI agents — product research, marketing automation, supply chain optimization, and business analytics for online sellers across Amazon, Shopify, Etsy, TikTok Shop, and all platforms.
At a glance
- What is it?
- The repository ships prompt-style skill files for Amazon, Shopify, Etsy, TikTok Shop, eBay and Walmart sellers, installed through the skills CLI rather than pip. It is a knowledge pack, not a runtime, and that distinction decides whether it belongs in your setup.
- Who is it for?
- Adopt this if you already run Claude Code, Cursor, Codex or another skills-compatible agent and you want structured e-commerce reasoning without wiring up a single store API. Skip it if you need live order, inventory or ad data, because nothing in the repository fetches any.
- Can I use it commercially?
- Yes. MIT is a permissive licence: you can use, modify and sell software built on it, as long as you keep its copyright and licence notices.
- Is it still maintained?
- Yes. The repository last received commits 21 days ago.
- What is it written in?
- Mainly Python, according to GitHub's language statistics.
Answers come from the project's GitHub data, last synced on September 15, 2026, and from our analysis. They are not legal advice.
DEEP OPEN-SOURCE ANALYSIS
What a Markdown Skill File Replaces
Most sellers who ask a general assistant about pricing get a generic answer, because the model has no reason to apply unit economics, fee tables or a market-scoring rubric to the question. This repository addresses that gap by shipping instruction files that describe how to approach a specific e-commerce task. The README describes them as plain-text instruction files that give AI agents specialized expertise in selling online, and states there are no binaries, no API keys and no setup friction. That claim is worth taking literally: installing a skill changes what the agent reads at inference time, not what data it can reach. So the product is a body of procedure, not a connector. If your problem is that your assistant gives shallow advice about FBA fees or international VAT, a skill file can fix it. If your problem is that your assistant cannot see your inventory, this repository does nothing for you.
The Folder Layout Behind the Skill Catalogue
The collection is organised by domain rather than by platform. The README lists Competitor Analysis, Pricing & Profitability and Growth & Expansion as the opening groups, with skills such as competitor-price-analysis, profit-margin-calculator-amazon and cross-border-ecommerce filed under them. Platform-specific logic lives inside individual skills instead: there are separate profit calculators for Amazon, Shopify, TikTok Shop and Walmart, each described with its own fee vocabulary (referral fees and FBA for Amazon, ad spend and CAC for Shopify, affiliate commissions and FBT for TikTok Shop, WFS fulfillment for Walmart). Each skill appears to be a directory containing a SKILL.md, which is the convention the skills CLI expects. The README marks entries with a status key, Available or Beta, and the pricing and competitor groups are listed as Available. That is a self-reported label, not a verification step, so treat Beta entries as documents still being revised.
Installing Through the skills CLI, Not pip
Python is the primary language of the repository, but installation does not go through pip. The README gives the skills CLI as the entry point. To install the whole collection globally:
npx skills add nexscope-ai/eCommerce-Skills -g
To install one skill, the README shows a --skill flag with the skill name:
npx skills add nexscope-ai/eCommerce-Skills --skill ecommerce-growth-strategy -g
After that, the documented workflow is conversational. The README's example prompt is a Shopify seller doing 8K a month asking how to reach 20K in six months, and the expectation is that the installed skill supplies the framework the agent reasons with. One inconsistency is visible in the README itself: the banner and the Quick Start text say 157 skills, while the install instruction says 142. The repository has no releases, so there is no changelog to reconcile the two numbers against. Check the skill count in your installed tree rather than trusting either figure.
Where the Skills Stop and Your Data Has to Start
The important limitation is structural. These files contain frameworks, not data. A profit calculator skill can walk an agent through a fee breakdown, but the referral percentage, the storage rate and the return rate have to come from you or from the conversation. Nothing in the material suggests any skill authenticates against Amazon SP-API, Shopify Admin or TikTok Shop. That has two consequences. First, any number the agent produces is only as current as what you typed. Second, the review-checker skills, which the README describes as analysing fake reviews through time clustering and verified purchase validation, cannot inspect a live listing on their own; they describe what to look for when review data is supplied. If your workflow depends on scheduled pulls or automatic repricing, this repository is the wrong layer, and the dynamic-pricing-ecommerce skill is a rule-design document rather than a repricing engine.
A Different Shape of Tool: Sellerboard and Similar Analytics Suites
The nearest alternative for the profitability side is a hosted seller analytics suite such as Sellerboard, which connects to your Amazon account and computes profit per SKU from actual orders and fee events. The difference is not quality, it is where the logic runs. Sellerboard ingests your transaction data and produces figures; this repository produces reasoning about figures you supply. That makes the hosted suite better for monitoring and reconciliation, and this repository better for planning conversations where the input is a scenario rather than a ledger. A seller who wants a defensible profit number for last month should use the connected tool. A seller who wants to think through whether a price change is worth testing can use the price-optimization-tool skill, which the README describes as covering unit economics, elasticity, candidate scenarios, experiment design and rollout gates. The two are not substitutes, and running both is normal.
Licence, Forking and What Maintenance Looks Like
The repository is MIT licensed, which permits commercial use, modification and redistribution provided the copyright notice and permission notice are retained. That matters for agencies: you can fork the collection, edit the frameworks to match a client's fee structure, and keep the changes private, because MIT does not require you to publish derivatives. It does not grant rights to any third-party trademark or platform rule text the skills may quote, and fee schedules change often enough that a fork will drift. There are no releases, so upgrades are whatever lands on main. The last push recorded is 2026-08-26. Practically, that means pinning is manual: if you fork, record the commit you forked from, because npx skills add pulls from the default branch and a later pull can silently replace skill text you had adapted. This is not legal advice; read the LICENSE file in the repository before redistributing.
Who Should Install This and What to Read First
The README states the collection works with OpenClaw, Claude Code, Cursor, Windsurf, Codex and any Skills-compatible agent, so the prerequisite is an agent that reads skill directories. Within that constraint, the strongest fit is a seller or agency that already has store data in spreadsheets and wants the assistant to apply a consistent method to it: fee breakdowns, market scoring, listing copy, PPC budget splits. The weakest fit is anyone expecting automation, since no skill in the material executes a platform API call. The README also points to a separate Amazon-Skills repository with 51 Amazon-specific skills, which suggests overlap if you sell mainly on Amazon; compare the two before installing both. Start by reading the SKILL.md for the single skill closest to your bottleneck, run the documented install command for that one skill, and judge the output on a question you already know the right answer to.
Editorial conclusion
Adopt this if you already run Claude Code, Cursor, Codex or another skills-compatible agent and you want structured e-commerce reasoning without wiring up a single store API. Skip it if you need live order, inventory or ad data, because nothing in the repository fetches any. Before you commit, open the skill directory you actually care about, read its SKILL.md, and check whether the frameworks inside match how you sell; the README's own mismatch between 157 and 142 skills is a reminder that the index and the folders can drift apart.
Community notes