Terminal
Terminal/Skill
Finds the single real company website for a company name, or flags it as ambiguous, not found, or acquired, before you enrich a list.
By Clay
Works with
npx skills add clay-run/clay-skill-creator --skill resolve-company-domainThe insight: a wrong domain poisons every row downstream — refusing beats
guessing. The naive version takes the first search hit or whatever a lookup
returns; the failure is silent, and every enrichment, signal, and email built on it
inherits the wrong company. So this skill treats any looked-up domain as a
CANDIDATE, validates it actually belongs to the operating company (not a parent, a
brand redirect, or a similarly-named stranger), and returns ambiguous or
not_found — with candidates — when the name doesn't pin one entity. An honest
refusal costs a re-ask; a confident wrong domain costs the whole row, invisibly.
Nothing here ships with a value. Each one is the installer's, not the author's: ask for it, never substitute a plausible default, and where an answer does not exist say which step becomes unavailable rather than guessing. Where a default IS defensible it is named below, and using it means saying so in the output.
| Input | What the installer supplies | If it is missing |
|---|---|---|
| What they have per row | a name plus a claimed domain, a name only, a domain only, or a company profile URL | no default — each route costs differently, and a claimed domain is validated rather than looked up, which is free |
| Cost ceiling | credits | dedupe names first, then state lookups × declared cost plus validation per survivor, and wait |
If an answer sheet is present beside this skill, load it and ask only for what it does not cover. A partial sheet is normal; a value it is missing gets asked for on its own rather than restarting the interview. Say which values came from the sheet before using them — a sheet applied silently is a wrong field nobody catches. If there is no sheet, say nothing about sheets — the check is a file lookup, not a question, so run the interview as though the feature did not exist rather than reporting an absence. At delivery, offer to save the answers back (identifiers only — never a token or a password), private and never published — and phrase the offer so it explains itself: "want me to save your answers to a file, so the next person on your team doesn't have to answer these again?"
Run clay whoami; echo "exit_code=$?". If it fails or Clay tools are missing, run
the Clay plugin's setup skill, restart if it says to, and re-run this skill. Tell
the user which workspace you're in. Confirm the managed Company Domain function
exists and read its declared cost (clay routines list, then clay routines get <id> — get requires the routine id).
| You have | Path |
|---|---|
| Name + a claimed domain | validate the claim (Step 3) — no lookup spend |
| Name only | candidate lookup (Step 2) → validate |
| Domain only | validate (Step 3), entity name comes from the site |
| Neither, but a LinkedIn company URL | pull the site link from the profile → validate |
Batch input: dedupe names first; state cost (lookups × declared cost + ~0-1 credit validation per survivor) and get approval before running.
Run the managed Company Domain function (name → domain, ~1 credit). Its output is a CANDIDATE, never a result — the function resolves a company for the name, not necessarily the company (name collisions are the #1 failure), and lookups on ambiguous or common-word names return confident wrong answers. Optional second candidate source when hints exist (a known country/region, industry): the company search arm — noting its identifier filter is recall-not-exact and can miss the canonical entity entirely (gate on match confidence, never take position as truth).
Run the ladder in references/validation-ladder.md — in order, cheap first:
not_found evidence; a redirect landing on social media
or a parking page → inactive candidate; a redirect to ANOTHER domain → follow it
and validate the destination (brand → corporate redirects are common).acquired with both
the stale domain and the acquirer's domain named; a REBRAND of the same entity
(same company, new name/site) may still resolve, with the reasoning stated.
Enrichment corroboration when needed (~1 credit): the payload's website field
(never its domain field, which can echo a link-shortener) — and remember
enrichment PRESENCE proves the entity exists in data, never that the domain is
alive: dead and acquired companies enrich fine on last-known data.canonical_domain +
operating_entity_name + confidence (validated / corroborated) + provenance
(which gates it passed, quoting evidence).acquired + the acquirer's domain as the actionable candidate
(resolving to the acquirer is a USER decision — the entity changed).website, never domain; MUST apply
name-boundary discipline to candidate entities.acquired
verdict exists for exactly this); NEVER let enrichment presence stand in for
liveness — dead companies enrich fine on last-known data; only the probe answers
"is this domain alive".Ask: "Clean these 5 company names into real domains: Brightloop, Meridian, Subway, Quartzlane Systems, Zzyqx Dynamics."