Let Claude Code or Cursor set up your page
Coding agents do not need a tool server to work with Tapmy. They need a procedure, and the procedure fits in one file.
What a skill is
A skill is a Markdown file with a short header — a name and a one-line description of when to use it — followed by instructions. Claude Code, Codex, Cursor and similar agents read the description of every installed skill and pull in the full text when a request matches. Ours matches “link in bio”, “links page”, “Linktree alternative” and the like.
Installing it
- Claude Code: `mkdir -p ~/.claude/skills/tapmy-link && curl -s https://tapmy.link/skill.md -o ~/.claude/skills/tapmy-link/SKILL.md`.
- Cursor and others: save the same file where the agent looks for skills or rules, or paste its contents into a project rule.
- No API key is needed to start. The agent gets one when it creates the first page.
What it makes the agent do
The procedure is deliberately boring. Ask the person for what only they know — address, name, bio, links, socials, a photo, anything 18+. Check the address. Validate the whole request with the dry-run endpoint so every problem shows up at once. Create the page with an idempotency key, so a retry cannot make two. Read the page back as text and compare it with what was asked. Then hand over the address and the claim link.
It also carries guardrails: one page unless asked for more, mark adult content, never publish anything the person did not ask for, keep the claim link private, send a User-Agent that names the tool. An agent following the file behaves like a careful assistant, not like a script.
Keeping the agent on the page
The create response contains an API key. If you want the agent to maintain the page — add this week’s link, swap a template, report statistics — tell it to keep the key in its environment. Everything after creation is a `PATCH` with the key, and the docs list the fields. If you would rather it did not, revoke the key in your account settings after you claim the account.