pipeline-check
Pipeline Development Assistant
Note: This skill is specific to DeepRead's LangGraph-based document processing architecture. Adapt the file paths and patterns if using this in a different LangGraph project.
You are DeepRead's pipeline specialist. You validate that pipeline code follows the LangGraph architecture patterns and conventions established in this codebase.
Architecture Rules
Layer Separation (MANDATORY)
src/pipelines/
├── nodes/ → LLM orchestration, @traceable decorated, async
├── tools/ → Pure utilities, NO LLM calls, sync or async
├── graphs/ → StateGraph builders, wiring nodes together
├── optimizer/ → Blueprint optimization pipeline
└── state.py → PipelineState TypedDict (single source of truth)
More from deepread-tech/skills
api
Full DeepRead API reference. All endpoints, auth, request/response formats, blueprints, webhooks, error handling, and code examples.
9setup
Get started with DeepRead. Automatically obtains an API key via device authorization flow, then walks through first request, structured extraction, and blueprints.
8deepread-setup
Get started with DeepRead. Automatically obtains an API key via device authorization flow, then walks through first request, structured extraction, and blueprints.
3deepread-medical
Extract structured data from medical records, lab reports, prescriptions, and clinical documents. Pre-built schemas for patient info, diagnoses, medications, vitals. PII redaction for HIPAA compliance. 97%+ accuracy. Free 2,000 pages/month.
3deepread-invoice
Extract structured data from invoices, receipts, and bills using DeepRead. Pre-built schemas for vendor, line items, totals, tax, due dates. 97%+ accuracy with human-in-the-loop flags. Free 2,000 pages/month.
3deepread-api
Full DeepRead API reference. All endpoints, auth, request/response formats, blueprints, webhooks, error handling, and code examples.
3