What problem does it solve? Before committing to an architecture, product teams need to know which third-party integrations their domain requires, but ad-hoc searching produces unfalsifiable claims with no recorded denominators, no verbatim queries, and no way to distinguish a service no catalog carries from a search that never ran. ## Core Features & Use Cases - Four-wave survey pipeline: builds an integration vocabulary map (wave 0), executes one of eight search angles across connector catalogs, OpenAPI descriptors, SDK registries, webhook conventions, and MCP channels (wave 1), deep-reads one admitted service into an extract record (wave 2), and synthesizes everything into an integration register through eight lenses (wave 3). - Deterministic schema gate: every artifact is validated by a Python script against JSON schemas, with cross-checks joining the register, extract records, and frozen queue in both directions. - Honest absence recording: zero-hit cells carry their queries and receipts, out-of-vocabulary values map to null or unknown rather than forced guesses, and injected instructions in third-party descriptor text are neutralized and recorded as a posture finding. - Use Case: Before designing a scheduling product's architecture, run the survey to produce an integration register ranking services like calendar and payments providers by catalog presence, directory hits, availability, and an auditable complexity score. ## Quick Start Use the integrations-prior-art-survey skill to map the integration surface for my product scope and produce a validated integration register.
