Run a per-user Telegram symptom follow-up workflow for genetic-risk screening. Use when a Telegram user sends /start for the symptom bot, when onboarding or restarting a user into the workflow, when asking or recording Round 1–5 answers, when handling emergency red-flag symptom replies, when creating or continuing per-user schedules/state under data/symptom-users, or when concluding after Round 5 and finding Vienna/Austria specialists on DocFinder. Also use when a cron or scheduled job sends the next round follow-up, updates data/symptom-users state, or logs a dashboard summary for a symptom screening run.
Run the Telegram symptom workflow as a per-user, per-chat longitudinal screening flow.
/start, ask for:
data/symptom-users/<sender_id>.json.The live agent uses cron (or equivalent schedulers) to remind users between rounds. When a job runs for a user who is still in the workflow:
data/symptom-users/<sender_id>.json: currentRound, per-round status (e.g. pending), askedAt, and the outbound Telegram messageId when available.If the environment commits workspace changes, record a clear commit message for the state update. See openclaw/README.md for an example dashboard log of a cron-driven Round 2 send.
Use the same core questions each round:
Rules:
Immediately advise urgent/emergency care if the user reports:
When emergency symptoms appear:
Do not diagnose. Use language like:
After 5 rounds:
If the pattern is strong enough after Round 5:
Maintain per-user state in data/symptom-users/<sender_id>.json with at least:
currentRoundmaxRoundscredentialsProvidedwebsiterounds[] with answers/status/timestampsdoctorSearchCompletedIf the reminder DB exists, keep it in sync with the state file.
references/workflow.mdexamples.md