メニュー
Extract, parse, and load documents from a ZIP archive or single file into SQLite + ChromaDB. Use when the user asks to "load", "ingest", or "import" a document.
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
Orchestrate a full multi-agent document review pipeline in one pass — load, evaluate, verify, draft, and audit.
Display the audit trail of all document analysis actions. Use when the user asks to see the "audit trail", "history", or "what we've done".
Session timeline viewer and manual entry recorder for the audit trail.
Evaluate loaded documents against a criteria file, producing a structured findings report. Use when the user asks to "evaluate", "review", or "check" documents against criteria.
Use macOS say command for audio notifications when user input is needed, a long task completes, or an error occurs.
Search loaded document clauses using semantic and keyword search. Use when the user asks to "search", "find", or "look up" something in the documents.
| name | load-document |
| description | Extract, parse, and load documents from a ZIP archive or single file into SQLite + ChromaDB. Use when the user asks to "load", "ingest", or "import" a document. |
| argument-hint | <path-to-archive-or-file> |
Extract, parse, and load documents from a ZIP archive (or single file) into SQLite + ChromaDB.
If the path is a ZIP archive, extract all documents first.
Run the loader to parse and populate both data stores:
uv run --with 'pymupdf,chromadb,sentence-transformers,python-docx,openpyxl' \
"${CLAUDE_PLUGIN_ROOT}/skills/load-document/document-load.py" "$ARGUMENTS"
Report results to the user:
If no path is given, use: assets/1-RFP 20-020 - Original Documents.zip
uv run --with to manage dependencies — never pip install