01Utilize TOON-encoded responses and compact data formats to significantly reduce token usage for LLMs.
020 GitHub stars
03Enforce daily budget limits for various X actions (replies, posts, likes, retweets, follows).
04Implement destructive tool gating, hiding dangerous operations like 'delete_tweet' by default.
05Prevent duplicate engagement with tweets, tracking past interactions to avoid spamming.
06Offer client-side engagement filtering and relevancy sorting for `search_tweets` to surface high-quality results.