X Timeline Digest

**Quick judgment** A well-scoped, practical skill for converting noisy X/Twitter timelines into a deduplicated, high-signal JSON digest, optimized for downstream Chinese-language summarization. It’s best suited for users who already rely on the bird CLI and want to integrate X updates into automations or notification channels. **Strengths** - Uses both For You and Following timelines for richer coverage. - Incremental, stateful processing (lastRunAt + sentTweetIds) to avoid re-sending tweets. - Hard and near-duplicate deduplication plus heuristic spam/ad filtering. - Clean, well-documented JSON output tailored for LLM-based digest generation. **Key risks & limitations** - **External dependency fragility:** Relies on bird and X’s underlying behavior; any changes to X or bird can break fetching. - **Auth & environment setup:** Requires bird to be installed, in PATH, and authenticated via cookies; may be non-trivial for less technical users. - **State file robustness:** Corruption or deletion of the state file can cause missed or duplicated items; no explicit resilience mechanisms are described. - **Content safety:** It surfaces whatever is in the user’s timelines; harmful or NSFW content can be included unless filtered in upstream or downstream workflows. - **Language targeting:** The recommended smart-digest flow assumes Simplified Chinese output, which may not fit all audiences. **Recommended scenarios** - You already use bird and want a **repeatable, automated digest** of your X/Twitter activity for personal or team use. - You’re building an **LLM-powered “smart brief”** (especially in Chinese) on top of your timelines, and need a clean JSON feed to summarize. - You want to **reduce noise** (ads, low-value tweets, near-duplicates) before sending X updates to Telegram, email, or other channels via OpenClaw.