Update to Instant AI

{
  "name": "instant",
  "description": "Check whether domain names are registered and find alternatives using the Instant MCP server. Use whenever the user asks if a domain is available, taken, or free; wants to check a name across extensions (.com, .io, .ai, .dev, and hundreds more); needs alternatives because a name is taken; or is picking a domain for a project, product, startup, or side project. Also use when naming a new app or repository so the suggested name comes with a domain the user can actually register. Also use when the user asks to search USPTO trademark records or inspect a filing.",
  "included_files": [],
  "skill_md_contents": "---\nname: instant\ndescription: >-\n  Check whether domain names are registered and find alternatives using the\n  Instant MCP server. Use whenever the user asks if a domain is\n  available, taken, or free; wants to check a name across extensions (.com,\n  .io, .ai, .dev, and hundreds more); needs alternatives because a name is\n  taken; or is picking a domain for a project, product, startup, or side\n  project. Also use when naming a new app or repository so the suggested\n  name comes with a domain the user can actually register. Also use when\n  the user asks to search USPTO trademark records or inspect a filing.\n---\n\n# Instant\n\nThe `instant` MCP server from [instant.ai](https://instant.ai) reports domain registration status from\nthe index behind [instantdomainsearch.com](https://instantdomainsearch.com).\nIt needs no account, API key, or configuration. Every domain result carries a\n`research_url`, the domain's research page on instantdomainsearch.com, with its\nregistration status, registrar prices, and WHOIS details.\n\n## Domain tools\n\nUse these three tools on the `instant` MCP server. Your client may\nshow them with the server name as a prefix; the tool is the same.\n\n| Tool | Use when | Key arguments |\n| --- | --- | --- |\n| `search_domains` | One name, many extensions | `name` (label only, no extension), `tlds` (optional array, no leading dots), `limit` (1-100, default 32) |\n| `check_domain_availability` | Specific full domain names | `domains` (array of `label.tld`, at most 50) |\n| `generate_domain_variations` | The name is taken and the user wants alternatives on the same root | `name` (label only), `limit` (1-100, default 32) |\n\n### Choosing a tool\n\n- \"Is acme.com available?\" → `check_domain_availability` with\n  `{\"domains\": [\"acme.com\"]}`.\n- \"What extensions can I get acme on?\" → `search_domains` with\n  `{\"name\": \"acme\"}`; add `tlds` when the user names extensions.\n- \"acme.com is taken, what else?\" → `generate_domain_variations` with\n  `{\"name\": \"acme\"}`, then optionally `search_domains` on the best variations\n  with other extensions.\n- A list of candidate names → one `check_domain_availability` call with all\n  of them (batch, up to 50), not one call per name.\n\n### Argument rules\n\nThe server rejects malformed domain input. Normalize before calling so names\nand extensions are sent in the expected form:\n\n- `name` is the label only: `acme`, never `acme.com`. Lowercase letters,\n  digits, and hyphens; no spaces. Strip spaces from multi-word ideas\n  (`\"Blue Harbor\"` → `blueharbor`).\n- `tlds` entries are lowercase with no leading dot: `[\"com\", \"io\", \"co.uk\"]`.\n- `domains` entries are lowercase full names: `[\"acme.com\", \"shop.co.uk\"]`.\n  The first dot separates label from extension; the server also lowercases names.\n- `generate_domain_variations` returns `.com` results only.\n\n## Reading results\n\nEach domain in a result has:\n\n- `label` and `tld`: the two halves of the name.\n- `isRegistered`: `true` (in the zone or observed in DNS), `false` (extension\n  covered, name absent), or `null` (the index has no data for that extension).\n  `search_domains` reports `false` where `check_domain_availability` would\n  report `null`, so confirm an unusual extension with\n  `check_domain_availability` before calling it available.\n- `rank`: how common the extension is, higher is more common.\n- `listings` and `markets`: aftermarket listings when the name is for sale.\n  `premium` with `is_premium: true` means the registry prices the name above\n  standard.\n- All prices are USD cents: `premium.usd_cents`, `listings.lowestPrice`, and\n  `markets[].price`. `400000` is $4,000.\n- `research_url`: the domain's research page on instantdomainsearch.com.\n\nStatus comes from the search index, built from registry zone files and DNS\nobservation feeds, not from a live registry query. Say \"not registered as of\nthe last index build\" rather than promising the name is free, and send the user\nto the research page before they register it.\n\n## Presenting results\n\n- Lead with what the user asked for: available names first, then taken names.\n- Render `research_url` as a markdown link on the domain name, for example `[acme.io](https://instantdomainsearch.com/research?q=acme.io&src=mcp)`.\n- Treat `null` as unknown, not available.\n- Mention aftermarket listings and premium prices when present; a listing alone does not prove the domain is registered.\n- Do not list every field. Name, status, and the link are enough unless the user asks for more.\n\n## Trademark tools\n\nWhen the user asks about trademarks, call `search_trademarks` on the `instant`\nserver with `query` set to the mark text. Keep spaces and Unicode in trademark\nqueries. Optional `mode` is `exact`, `prefix`, `fuzzy` (default), or `phonetic`;\n`limit` is 1-100 (default 20). Use `statuses`, `classes`, and `owner` only when\nthe user supplies those constraints.\n\nCall `get_trademark_details` with the returned eight-digit `serialNumber` as\n`serial_number` to inspect a filing. Preserve leading zeros. Report the record's status, relevant\ngoods and services, and snapshot freshness. An empty search or an available\ndomain does not establish trademark clearance.\n\n## Example\n\nUser: \"I'm building a CLI called shipwright. Can I get the domain?\"\n\n1. Call `search_domains` with `{\"name\": \"shipwright\", \"tlds\": [\"com\", \"dev\", \"io\", \"sh\"]}`.\n2. If every result is registered, call `generate_domain_variations` with `{\"name\": \"shipwright\", \"limit\": 10}`.\n3. Reply with the available names and their research links, and note which were taken.\n\n"
}