timelinesai-claude-code-skills

Use when the user asks analytical questions about their WhatsApp workspace — response time, unanswered chats, unread queues, volume or workload, topic trends, complaint tracking, lead scoring, campaign attribution, or any open-ended "why/what/who/how many" question about WhatsApp customer conversations. Fetches live data from the TimelinesAI public API ($TIMELINES_AI_API_KEY) and aggregates client-side. Trigger words include WhatsApp, TimelinesAI, first-reply time, SLA, unanswered, inbox, unread, response time, lead, qualified, refund, complaint, and any mention of specific customers by name or phone.

Use when the user wants you to draft WhatsApp reply messages WITHOUT actually sending them. Saves each draft as a private TimelinesAI note (POST /chats/{id}/notes) prefixed with "DRAFT:" so humans review and send manually from the TimelinesAI inbox. Trigger phrases include "draft replies", "draft me", "don't send", "review before sending", "suggest replies", or any request for WhatsApp suggestions where the user clearly wants human review in the loop. This skill deliberately does NOT call POST /chats/{id}/messages — use the broader timelinesai-whatsapp-analytics skill for real sends.

Use when the user reports that something looks wrong with their TimelinesAI setup — Claude Code says the token doesn't work, a WhatsApp number has gone quiet, a webhook stopped firing, or they just installed the skills and want to confirm everything is wired up. Runs smoke tests against the TimelinesAI public API and reports what's healthy vs what's broken. Trigger phrases: "token isn't working", "why can't you see my numbers", "setup check", "is the token correct", "verify my TimelinesAI connection", "smoke test", "quota left", "is the webhook still registered".