social-community-engagement

To automate drafting social media replies, provide a thread context and the system classifies it as a question, complaint, praise, or troll, then generates context-aware drafts with tone variants. Drafts are saved to workspace/drafts/social/replies/ for human review before sending.

No, classified social replies cannot be sent automatically by default. A human sends all outbound messages unless automation is explicitly approved in USER.md or TOOLS.md configuration files.

Escalation routing for social media management works by classifying thread severity into low, medium, or high levels and applying policy checks. The system logs handoff decisions and routes high-severity items like complaints for explicit human review.

Yes, the social reply tool handles public comments differently than direct messages by applying distinct DM vs public channel handling rules. This ensures appropriate tone variants and escalation decisions are generated for each specific social channel format.

The best way to classify social media threads involving trolls and complaints is using a multi-step workflow that categorizes intent, assigns severity, and applies policy checks. This process generates safe response drafts and explicit handoff rules for troll interactions.

To generate multiple tone variants for safe social media responses, input the thread context and the system drafts different response options based on classification. These variants are output to workspace/drafts/social/replies/ with standardized naming for easy selection.