Agents

Skills

A skill is a SKILL.md file an agent reads before it starts a job: when to use AffiliateSpy, the exact tool calls in order, what is metered, and what needs the user's yes. Three are published, generated from the live tool catalog so they never describe a tool the server does not have.

The skills

find-creators-promoting-competitor-app

Find, grade and contact the TikTok, YouTube and Instagram creators already selling a competitor app, then draft the campaign.

recruit-affiliates-already-promoting-competitors

Grow an affiliate or partner program by recruiting the sites and creators already earning from competitors, with pipeline and Autopilot.

find-best-of-roundups-for-your-app

Find the roundup and review pages ranking for your keywords, where competitors are listed and you are not, and reach the writer.

Index for tooling: /skills/index.json. Source and issues: github.com/Wayfinity/affiliatespy-skills.

Install

Claude Code reads skills from a folder. Download one and drop it in:

mkdir -p .claude/skills/find-creators-promoting-competitor-app
curl -o .claude/skills/find-creators-promoting-competitor-app/SKILL.md \
  https://www.affiliatespy.io/skills/find-creators-promoting-competitor-app/SKILL.md

Use ~/.claude/skills/ instead of .claude/skills/ to install for every project. Codex, Cursor and other harnesses that read SKILL.md files take the same folder; any agent can also fetch the raw URL at run time. The skill expects the MCP server or REST API to be connected first (see the quickstart).

How they behave

Every skill follows the same rules: read get_account first, present evidence rather than invent metrics, say what a reveal will cost before spending it, and never call a guarded tool (launch, send, scan, autopilot changes) without an explicit yes from the user in the conversation. The skills are regenerated whenever the tool catalog changes, so the tool names and arguments in them always match /openapi.json.