What problem does it solve? Recruiting searches often start from a company set rather than a person, and translating fuzzy requests like "series B fintech startups" or "companies backed by Sequoia" into concrete company identifiers is error-prone. This Skill converts company names, semantic descriptions, investor filters, and vertical queries into canonical TurboPuffer company IDs that downstream people search can consume. ## Core Features & Use Cases - Company Resolution: Runs the resolve_companies primitive against TurboPuffer to turn names, semantic queries, sector types, funding, headcount, and geography filters into canonical company IDs. - Investor Resolution: Resolves investor_names into investor URNs first when funding-source filters are part of the request. - Set Scoping: Resolves operator scope via resolve_set_operators, inheriting the default set ID or falling back to the logged-in operator's active personal set. - Sector Strategy Control: Supports staged, soft_union, hard_filter, and semantic_only sector strategies to balance precision and recall for domain searches. - Use Case: A recruiter asks for "database companies" — the Skill builds a semantic query with a staged sector strategy, resolves the company set, and hands the resulting company_ids to the people-search skill. ## Quick Start Ask the agent to run /search-company with a query such as "series b fintech startups in california" to resolve the matching company set.
