Six ready-made workflows for AI assistants that are connected to the Spytrend MCP server. A skill is
a plain folder: SKILL.md tells the assistant when the workflow applies and how to run it, agents/
carries the Codex descriptor, references/ holds the report template. The assistant reads the skill,
decides which tools to call, and renders a report you can send on.
| Skill | Ask it when you want to |
|---|---|
spytrend-vertical-intel |
map one advertising vertical in one market: who leads, which networks run it, what is being reused and launched |
spytrend-advertiser-overview |
profile a single advertiser: pages, countries, categories, what it runs now |
spytrend-webmaster-overview |
see everything the account behind a group of ads runs, across pages, domains and countries |
spytrend-website-review |
review a website as a traffic surface: who sends traffic to it and how it is advertised |
spytrend-keyword-search |
find what is advertised for a keyword in a market and read the result as a report |
spytrend-shop-discovery |
discover shops in a niche and market with their traffic and advertising signals |
Claude Code (plugin, one command):
/plugin marketplace add Spytrend/spytrend-mcp
/plugin install spytrend-mcp-skills@spytrend-mcp
Any assistant via the skills installer (Claude Code, Codex, Cursor, Copilot, Gemini CLI and 70+ more):
npx skills add Spytrend/spytrend-mcpClaude Code. Copy the skill folders into your skills directory and start a new session:
cp -R skills/spytrend-* ~/.claude/skills/Codex. Each folder carries agents/openai.yaml; point Codex at the folder or copy it into your
Codex skills directory.
Automatic discovery. The server publishes the same pack at
https://mcp.spytrend.com/.well-known/agent-skills/index.json; assistants that read that index pick
the skills up without a copy step.
The skills call the Spytrend MCP server — connect it too. The plugin route wires the server
for you. If you installed the skills with npx skills add or by copying the folders, add the
server separately: point your client at https://mcp.spytrend.com/mcp and sign in with your
Spytrend account in the browser. Full per-client instructions are in the
repository README.
contract.md is the shared contract: how arguments are read, how cost is checked with
get_usage before anything is spent, what a report must contain and how it is rendered
(scripts/render_report.py, templates under each skill's references/). manifest.json lists the
skills, the tools each one calls and the input schema (schemas/) it validates its arguments against.
The live tool list and the fields each call returns are authoritative when they differ from a cached client description — every skill says so and is written to re-check rather than assume.