What problem does it solve? Staying current on AI news, papers, and projects from X usually means doom-scrolling or paying for an expensive X API tier. This Skill scrapes public X profile pages through the Firecrawl CLI and compiles recent posts from a curated list of high-signal accounts into a single local HTML page you can browse at your own pace. ## Core Features & Use Cases - Curated source list: A hand-editable sources.json maps X handles to categories (Labs, Research & Voices, Agents & Tooling, Builders), which drive filter tabs and badges on the generated page. - Self-contained HTML output: scripts/feed.py scrapes each handle, filters posts by a configurable lookback window, deduplicates, and renders a standalone feed.html with search, category tabs, and a headline ticker. - Cost-aware operation: Each run costs one Firecrawl credit per handle, with retry logic and per-handle error reporting so you know which sources came back empty. - Use Case: Ask to refresh the AI feed each morning; the Skill scrapes 17 curated accounts, keeps posts from the last 24 hours, and opens a local page you can filter by category or search by keyword. ## Quick Start Ask the assistant to refresh the AI feed with a 24-hour lookback and then open the generated feed.html page.
