Expand Reputelix Social ORM

I have already released the core of Reputelix—monitoring, complaint management, workflow, reporting, and an intelligence dashboard are in place and running smoothly. What I need now is help finishing the social-facing pieces so brands can act on feedback the moment it appears online. Social media sources The platform must ingest public posts, comments, and direct messages from Facebook, Twitter, and Instagram in near-real-time. Robust scheduling, rate-limit awareness, and retry logic are essential so nothing slips through when APIs change or throttle. User authentication End-users should sign in with their existing social media accounts. Please wire up OAuth2 flows for Facebook Login, “Sign in with Twitter,” and Instagram authorization, then hand back the tokens to our existing JWT-based auth service. Session refresh and revocation endpoints must also be exposed. Issue classification Every incoming mention needs to be auto-tagged as a Customer service complaint, Product defect, or General inquiry. I’m open to the approach—lightweight rules engine or supervised model—as long as it returns a label and confidence score we can surface on the dashboard. The classifier should plug into our case-creation queue so the right workflow is triggered immediately. How it plugs in All code must be container-ready (Docker) and fit our current stack (Node.js/TypeScript backend, React frontend, PostgreSQL). I’ll grant repository access and share the existing GraphQL schema you’ll extend. Clear README notes and postman collections for every endpoint are mandatory so my QA team can verify everything quickly. Deliverables • API connectors for Facebook, Twitter, Instagram with incremental sync • OAuth2 social login flow fully integrated and tested • Classification service (source code + model) returning label & confidence • Updated GraphQL mutations/queries and React components where needed • Dockerfile(s) and setup scripts, plus concise deployment docs Acceptance criteria 1. A new social post reaches the dashboard within 60 seconds of appearing on any monitored platform. 2. A first-time user can create an account by authenticating with one of the three networks and receive a valid JWT. 3. The classifier labels test data with ≥85 % accuracy across the three categories. 4. All new services pass the existing CI suite and expose health checks for Kubernetes. If you have recent experience with social platform APIs, OAuth2, and real-time data pipelines, I’d love to see how you can push Reputelix to the next level.