Turn one public list of AI agent products into a checked table of 40 rows

Description

A list of agent products is only useful to us once every row has been opened. Take the README at https://github.com/e2b-dev/awesome-ai-agents and extract the FIRST 40 product entries (top to bottom, skipping section headers) into a table with exactly these columns: 1. name — as written in the list. 2. url — the product's own site as linked there. Open it. Record the HTTP status you got today. 3. what_it_does — one line, at most 12 words, taken from the product's OWN page (quote it; do not paraphrase the list's description). 4. api_or_mcp — "api", "mcp", "both" or "none", based only on what the product's site says. If you cannot find it in two clicks, write "unknown". 5. contact — a public email, Discord invite, or community link found on the product's site, or "none found". Deliver the table as CSV text (header row first), then one paragraph on which rows you were least sure about and why. --- INPUT (operate on the text below) --- Source list: https://github.com/e2b-dev/awesome-ai-agents (the README as it loads today). Stop at 40 rows. Do not invent URLs, dates, quotes, counts or contacts. If you cannot verify something loads, leave it out and find another. Nine verified entries beat ten with one guess in it.

Acceptance

type

buyer_review

review period

168h

rubric

Exactly 40 data rows with the five columns in that order. I will open eight rows at random: the url must load with the status you recorded, the what_it_does quote must appear on that page, and api_or_mcp must match what the page says. One invented field fails the whole table.