Collect public X posts from logged-in Chrome DOM into normalized JSONL.
npx skills add https://github.com/binggandata/bggg-skills --skill bggg-data-x
Please help me install this Agent Skill. Skill: bggg-data-x Source: https://github.com/binggandata/bggg-skills/tree/main/bggg-data-x Command: npx skills add https://github.com/binggandata/bggg-skills --skill bggg-data-x
X/Twitter blocks most automated access behind login walls, paid APIs, and internal GraphQL endpoints, making it hard to gather auditable public post data for voice-of-customer research. This Skill collects visible posts through the user's own logged-in Chrome session and normalizes them into a consistent, auditable JSONL dataset.
Use bggg-data-x to search X in my logged-in Chrome for these keyword queries and save the collected posts as normalized JSONL with an audit summary.
High-intent search queries and answers about installing and using this skill.
Drive the user's already logged-in Chrome session, open X search URLs, and extract rendered tweet cards from the visible DOM using selectors like article[data-testid="tweet"]. Results are saved per query and normalized into JSONL without touching internal GraphQL or cookies.
Build a query plan from a TSV file, open each generated search URL in one Chrome tab, validate the selector contract, then scroll and accumulate posts keyed by Tweet ID. Stop after six stale scrolls or a 180-scroll hard cap and save one source package per query.
No. The collection reads only visible rendered DOM content and explicitly forbids inspecting or exporting cookies, local storage, credentials, or profile data. Authentication relies entirely on the user's existing signed-in browser session.
Each row preserves the numeric Tweet ID, canonical status URL, timestamp, exact post text, author handle and display block, engagement aria-label, full card text, query, language hint, and collection time. Normalization also parses replies, reposts, likes, bookmarks, and views.
X search is a visible sample, not a complete census, so results depend on sort mode, query syntax, and date slicing. Collection stops on challenge pages, rate limits, or blank timelines, and search-engine snippets must never substitute for original post text.
OpenAnalystIncTrigger parallel research agents across 13 platforms for marketing intelligence.
mguozhenAnalyze Amazon product reviews to extract sentiment, pain points, and selling points.
Synter-Media-AIDeliverable: A unified statement of the requested format.
githubIntegrate Xquik APIs, SDKs, and Apify Actors for X data extraction workflows.
