メニュー
Extract an inventory of entities (person names, money, locations, organizations, time, dates) and extract main keywords from the provided text to describe the material.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
A skill to resolve Decentralized Identifiers (DIDs), search any information using ODRL services.
Save highlighted or provided text as a note in a local file with a unique name.
Find contradicting interpretations of a fragment of text, comparing pros and cons.
Resolve Decentralized Identifiers (DIDs) and search for information using ODRL services. Use when the user needs to fetch metadata for a DID or search for resources with DID identifiers.
Extract English transcripts from YouTube videos when requested.
Extract from text all quantitative and qualitative variables, numerical measurements, geolocation, shape, temperature, distance, and timespatial variables and make detailed description of all measurements with indication of units.
| name | inventory |
| description | Extract an inventory of entities (person names, money, locations, organizations, time, dates) and extract main keywords from the provided text to describe the material. |
When the user asks you to "make an inventory", "inventory this", or sends content with instructions to extract entities and keywords, you must instantly perform the following analysis:
Entity Extraction: Read the provided text or transcript and systematically identify and list all significant entities, categorized clearly into the following headers:
Keyword Extraction: Identify and extract the 5-10 most important "Main Keywords" or key phrases that best define the core themes and topics of the material.
Presentation: Output the results cleanly using standard Markdown formatting. Use bulleted lists. If a specific entity category has no factual matches in the generated text, explicitly write "None found" beneath that header rather than omitting the category entirely.
JSON-LD Caching: Immediately after presenting the markdown, silently convert the extracted entity inventory into a valid JSON-LD format document mapping the concepts properly (e.g., schema.org). You MUST use your file writing tools or run a bash command (cat << 'EOF' > ...) to save this raw JSON-LD data to a uniquely named file inside the cache/ directory (for example: cache/inventory_TIMESTAMP.jsonld).
Provide the requested markdown inventory and a brief confirmation that the JSON-LD file has been cached, without extra conversational filler.
