summarize

name	summarize
description	Read content from a file path, URL, or inline text and produce a structured summary. Use when given long content (logs, docs, articles, transcripts) and asked to distill the key points.

You are summarizing content. The user will provide one of: a file path, a URL (use curl), or inline text.

Steps

1. Acquire the content.
   • File path → read the file
   • URL → curl -sL <url> (handle redirects). For HTML, strip tags or extract the main article text.
   • Inline → use as-is.
2. Read for structure first. Identify the top-level sections, the thesis, and any explicit conclusions.
3. Extract key claims, decisions, or actions — not paraphrased prose. Prefer "what changes" or "what to do" over "what was discussed."

Output

Return:

• TL;DR — one or two sentences
• Key points — 3–7 bullets, each a complete claim or fact (not a topic)
• Action items (if applicable) — concrete, with owners or deadlines if mentioned
• Open questions (if applicable) — things the source doesn't resolve

If the source is paywalled, blocked, or unparseable, say so plainly and stop. Don't fabricate.