What problem does it solve? Long-form research documents produced by AI assistants are often information-dense but hard to read: unexplained acronyms, filler phrases, bullet soup, and no synthesis. This Skill applies a reader-first style rubric so research reports, market analyses, and strategic memos are structured for re-reading, skimming, and decision-making. ## Core Features & Use Cases - Reader-first rubric: Enforces a TL;DR of standalone findings, acronym expansion on first mention, a decision-oriented "So what" per section, tables for comparisons, and inline footnote citations. - Attended and unattended modes: Asks bundled scoping questions (depth, audience, decision orientation, sources, format) by default, or picks documented defaults when running unattended. - Source discipline and self-review: Requires a sourcing plan, tracked footnotes with fetch dates, a fabrication check, and a grep-based self-review loop that bans tease-forward filler phrases. - Use Case: Ask for a competitive landscape of UK revenue-cycle-management vendors and receive a dated markdown report in scratchpads/ with a TL;DR, comparison tables, cited sources, and open questions. ## Quick Start Ask the assistant to research a topic, for example: "Do a deep-dive research report on the London SMB acquisition market and save it as a markdown document."
