Original workflows for websites, images, presentations, and everything in motion.
Read, create, or edit Excel spreadsheets (.xlsx/.xlsm) — sheet data, formulas, styles, charts, multi-sheet workbooks — and bulk .csv/.tsv tables; use whenever a spreadsheet is the input or the deliverable (extract/analyze data, add columns/formulas/formatting/charts, clean messy tables, build from scratch), but not for Google Sheets API or Word/PDF/script outputs.
Create and edit spreadsheets with explicit assumptions, consistent formulas, formatting and recalculation checks.
Build a structured paper-reading card covering methods, evidence chains, limitations and testable research ideas.
Simulate evidence-grounded peer review with independent reviewer reports and prioritized major and minor concerns.
Draft user-facing release notes from a verified release range.
Frame social-science research questions, literature maps, evidence routes and defensible claims, then review novelty and feasibility.
Review manuscripts through distinct journal-fit, methodology, domain and counterargument perspectives, with a structured revision roadmap.
Polish or translate academic prose while preserving facts, terminology, numerical precision and the scope of the evidence.
Reconstruct a paper’s reasoning, experimental design and evidence limits, with worked examples for key formulas.
Draft Chinese invention patent and technical disclosure materials from research evidence, including claims and prior-art comparisons.
Route a Korean editing request through the shared pipeline at the selected level of detail.
Split project memory into ISO weeks and generate connected narrative chapters with a carry-forward between weeks.
Rebuild scanned Chinese books as structured Word documents using MinerU OCR, title mapping and proofreading against the original pages.
Revise formulaic prose while preserving the writer voice, facts, intent and protected material.
Diagnose and revise formulaic Korean prose through the shared humanize-korean workflow.
Write and edit Japanese work documents, reports, meeting notes, articles, and emails while preserving the source's facts and level of certainty.
Turn experiment images, voice notes and text into structured Markdown logs with YAML metadata and optional Obsidian or Feishu integration.
Inspect Korean writing patterns and show a small sample revision before a fuller editing pass.
Create and edit Obsidian Bases (.base files) with views, filters, formulas, and summaries.
Turn books, papers, slides, and notes into source-grounded learning materials, interactive HTML study pages, and Markdown references.
Fetch web pages and PDFs, then return clean Markdown, summaries, or cited extracts.
Make targeted edits in an explicitly selected draft file and verify the resulting text.
Inspect writing patterns without rewriting, with an optional deterministic local Node.js detector.
Form research questions, search and verify literature, synthesize evidence, and produce a referenced research report.