summarizing-meetingslisted

# Summarizing — Universal Summarization Meta-Skill (model-agnostic) **Purpose**: Transform raw source material — a **meeting transcript** OR an **article / paper / thread / blog post** — into a highly-detailed, structured summary. The skill keeps its original meeting superpower and generalizes it into a universal summarizer that adapts to *what* it is given and *what shape* you want back. > The directory is still named `summarizing-meetings` for back-compat; functionally this is a > universal summarization harness. Meetings remain a first-class, unchanged path. ## 0. Two orthogonal axes (read this first) This skill is parameterized along **two independent axes**. Pick one value on each: | Axis | Values | How chosen | |------|--------|-----------| | **Content class** (*what is the input?*) | `transcript` (meeting/call/standup/retro/discovery) · `document` (article/paper/thread/blog/post/lesson) | **auto-detected** in Step 0; override with `--content transcript\|document` | | **Output format** (*what shape do you want back?*) | `pyramid` (two-level Markdown note — the default) · `note-json` (a structured note object for a knowledge-base / wiki import step) | **defaults to `pyramid`**; opt in with `--emit note-json` (neutral `title`/`body` fields) or `--contract wiki` (same object with `title_ru`/`ru_body` compatibility keys) | Any combination is valid: meeting→pyramid (the classic v1.0 behavior, **unchanged**), meeting→note-json, document→pyramid, document→note-json. The