What problem does it solve? Reading X (Twitter) content programmatically is blocked by login walls — unauthenticated scrapers like firecrawl only receive a degraded logged-out page. This Skill reads tweets, threads, bookmarks, long-form articles, and user timelines through X's authenticated web GraphQL API using session cookies, returning structured JSON that downstream marketing workflows can consume directly. ## Core Features & Use Cases - Five URL shapes supported: single tweets, same-author thread unrolls, bookmark collections, long-form articles, and user timelines, each dispatched to the correct GraphQL endpoint. - Frozen JSON output contract: every fetch returns a canonical six-field shape (tweet_text, thread_unroll, embedded_links, media_urls, author, timestamp) that source-digestion workflows consume without re-parsing. - Security-hardened pipeline: terminal-escape sanitization strips OSC 52 and CSI injection sequences, output is isolated to a gitignored .mktg-x/ directory, and engagement metrics are deliberately omitted to keep ideation unbiased. - Use Case: A user pastes a viral thread URL and asks for a content breakdown. The Skill fetches the full same-author thread via authenticated GraphQL, writes sanitized JSON to .mktg-x/, and the agent extracts only the fields needed with jq. ## Quick Start Set the MKTG_X_AUTH_TOKEN and MKTG_X_CT0 environment variables from your browser cookies, then ask the agent to read a tweet or thread URL such as "pull this thread https://x.com/user/status/123".
