semantic-web-fetchlisted

# Semantic Web Fetch A skill for interpreting a page **in conformance with** the profile it declares. This is not a procedure manual; it is the definition of where the authority over meaning resides. The procedure follows from the principles, so prioritize honoring the principles. ## Why this is needed By default, when given a URL I read extracted text (Markdown-equivalent). That is usually enough to understand the data by contextual inference — but such understanding is probabilistic guesswork, not the meaning the author declared. Worse, the extracted form structurally loses affordance information: which state a link transitions to. A profile supplies both of these — the authority of meaning and the transitions — as a machine-readable contract. The `rel="profile"` declaration is itself the contract that says "this page conforms to this definition," and this skill puts that contract into effect. ## The three principles (the definitions to uphold) ### 1. Authority Locate the authority over meaning in the profile the document declares. Do not make my own contextual inference the authority. Rather than guessing "this is probably the creation year," receive the fact that a declared descriptor *defines* it as the creation year. ### 2. Binding Treat only elements that can be bound to declared descriptors as grounds for meaning. Match HTML elements (typically the `class` attribute, or the respective attributes of microformats / RDFa / microdata) to descriptor IDs in the profil