What problem does it solve? Growing an X (Twitter) account organically requires constant monitoring of trending topics, drafting on-voice posts and replies, and learning from human edits — a repetitive workflow this Skill operationalizes with strict rules so any Claude session can act as the growth engineer. ## Core Features & Use Cases - Engine cycle execution: Run light, full, or weekly engine cycles via bash scripts/run.sh to collect trend data, refresh rankings, and generate AI digests. - Draft creation with guardrails: Create voice-linted Typefully drafts through the local ops API at 127.0.0.1:4242, with publishing structurally blocked so a human always approves. - Engagement intelligence: Rank accounts and topics with bun scripts/engage.ts and tune topics, accounts, and voice configs based on engagement data. - Edit mining and learning: Mine human edits against data/draft-originals.jsonl to propose dated edit-lessons and log session work to docs/log/. - Use Case: A social media operator asks Claude to run a full engine cycle, review the top trending topics, and draft three reply variants into a trending conversation — all linted against the persona's voice rules and queued as drafts for human approval. ## Quick Start Ask Claude to run a light engine cycle and then draft a voice-linted reply to the top trending topic as a Typefully draft.
