glossary-extract

Glossary Extract

Scans a set of pages for domain-specific terms that are used repeatedly but never actually defined, ranks them by frequency, and drafts a definition for each from context, so you get a prioritized list of what to add to the glossary instead of starting from a blank page.

Inputs

• {{PAGES}}: One or more documentation pages, concatenated. If from multiple files, separate them with a line like "--- ---" so the output can cite where each term first appears.

How to apply

Substitute each {{PLACEHOLDER}} above with the user's actual content, then follow the prompt below exactly — including its output format and its stated limits.

---

You are building a glossary by finding the gaps in one: every domain- specific term used across PAGES that is never explicitly defined anywhere in them.

What counts as a term

A domain-specific noun, acronym, or short phrase specific to this product or its problem space — not a common English word, not a generic software term every reader already knows (API, endpoint, JSON don't count unless PAGES uses them in a specific, non-standard way). If in doubt whether a reader outside this product would already know the term, it probably doesn't belong in the glossary.

What counts as "defined"

The term is used with an explicit definition somewhere in PAGES — a sentence that states what it is, not just a sentence that uses it in context. "A subscription is a standing request to receive events matching a filter" is a definition. Five sentences that each use "subscription" in a different code sample are not.

If a term is defined once but used many more times, it does not belong in the output — it only qualifies if it is undefined everywhere it appears.

Method

1. Scan PAGES for candidate terms.
2. For each candidate, check every occurrence — if none of them define it, it's undefined.
3. Count occurrences (frequency) and note where it first appears.
4. Draft a definition using only what context across PAGES implies about the term — do not use outside knowledge of what a term like this "usually" means in other products.