视觉画布设计
canvas-design
基于设计原则创作静态视觉作品,用于海报与品牌图形。
local-mcp-pdf-to-sheet【文档与办公】Use when the user wants to pull data out of one or more PDFs (invoices, bank/credit-card statements, receipts, reports, tables) and put it into a spreadsheet. pdf_read is one of the most-used tools; this codifies the read-PDF → structure → write-Excel workflow. Powered by LMCP on macOS and Windows.
独立 GitHub 快照 · 自动分类 · 以原作者文档为准
Reading PDFs and turning them into structured data is one of the highest-volume things LMCP users do. This skill codifies the end-to-end flow: read the PDF(s) on the user's computer (macOS or Windows), extract the rows, and build an Excel sheet. The tools execute on the device; the extracted text goes to the user's AI provider as part of the conversation.
pdf_read(path=...), one call per file. (Find the files first with
file_search, or finder_search on macOS, if the user only describes them, e.g. "my March invoices".)excel_create(path="<out>.xlsx", headers=["Date","Description","Amount"], rows=[[...],[...]], confirm=true).
headers is required; rows is the full data. For edits to an existing sheet use
excel_write_cell(path, cell="B2", value, confirm=true); to read one back use excel_read(path).excel_create (one row per record) —
don't write cell-by-cell unless editing an existing file.excel_create/excel_write_cell
are preview/confirm-gated (pass confirm: true once the user approves the columns).Source column so each row traces back to its file.