DalgoT4D

Generate, update, or review Dalgo user-facing documentation. Use when the user asks to document a feature, write a docs page, update existing docs, or after a PR ships to refresh affected docs.

Draft a new Dalgo release notes entry on dalgo_docs from the latest release tags of DDP_backend and webapp_v2. Use when the user asks to write, publish, or update release notes after a release ships.

Use when implementing a feature plan end-to-end across the Dalgo backend (DDP_backend, Django) and frontend (webapp_v2, Next.js). Triggers on "execute this plan", "implement this feature", "build this from the plan", working from a plan.md / tasks.md.

Use when the user wants to create a Linear issue from a feature plan document (plan.md) generated by /engineering/plan-feature or /engineering/ship-feature. Triggers on: "create linear issue from plan", "post this to linear", "make an issue for this plan", "communicate this plan on linear".

Transform messy ideas into structured PRDs that get stakeholder alignment before engineering starts building. Use when: write prd, product requirements, requirements document, feature spec.

Writing rules for plans, specs, research notes, and any user-facing explanation in Dalgo. Use whenever producing a plan.md, research.md, spec.md, or explaining a feature/decision to the user — so the reader (a non-technical NGO program manager, a first-time engineer on the codebase, or a busy reviewer skimming on a phone) can follow without backfilling jargon. Triggered by the planning, spec-writing, and design-handoff commands; also invoke directly when the user says "explain in simple language", "give me an example", or asks you to rewrite something more plainly.

Apply Tal Raviv's technology philosophy when evaluating tools, building features, making product or architecture decisions, or explaining technical concepts. Use when the user asks to demystify, evaluate, prototype, or think critically about any technology decision.

Apply a combined UX design expert and NGO user evaluation lens to any UI component, screenshot, or interface description. Use when reviewing UI for usability, accessibility, and suitability for non-technical NGO users.

Work with Dalgo data ingestion — Airbyte sources, connections, stream catalogs, and sync history. Use when a user asks to sync a source, check whether a sync ran, or inspect what a connection syncs.

Route Dalgo MCP tool workflows. Use when a request involves Dalgo data, pipelines, syncs, transforms, dashboards, charts, or reports, to pick the right domain skill and stay on MCP tools.

Run, monitor, create, and manage Dalgo orchestration pipelines (Prefect deployments), including the trigger-poll-verify loop. Use whenever a user asks to run a pipeline or check whether one succeeded.

Build and run dbt transformations in Dalgo — the transform canvas, operations, model materialization, dbt runs, and publishing to git. Use when a user asks to run dbt, build or edit models, or modify the transform canvas.

Diagnose Dalgo failures — failed pipeline runs, sync errors, dbt errors — and handle notifications. Use when a user asks why something failed, what went wrong, or what their notifications say.

Create and manage Dalgo charts, dashboards, and reports. Use when a user asks to visualize data, build or modify a dashboard, or snapshot data into a report.

Explore the Dalgo data warehouse — schemas, tables, columns, row counts, and sample data. Use for questions about what data exists or what values a table contains.

When computing a value, status, or derived property of a domain entity (KPI, metric, alert, chart, dashboard, etc.) from outside that entity's own service module, ask the owning service instead of rebuilding the logic. Use when writing a feature that touches another domain's data, fixing a bug whose root cause is parallel implementations drifting, or reviewing code that copies SQL/query-builder patterns from another module.

Prefer typed Pydantic/Ninja schemas over generic dicts when passing structured data through Dalgo backend code. Use when writing or modifying functions whose parameters represent a known shape (configs, payloads, deserialized JSONField values, frontend payloads). Skip for genuinely free-form / variable-shape data.

Scaffold and ship a new HTTP endpoint in the Django/Ninja backend using the API → Core → Schema → Exceptions layout. Use when adding a new route, creating a new module's routes, or extending an existing router with another verb.

Python coding standards for the DDP_backend covering service-layer function signatures (payload + orguser shape), exception chaining with `raise ... from err`, and import placement (top-of-file only). Use when writing or refactoring service-layer code, reviewing a diff or PR, asking "is this the right pattern", or finalizing code.

Write and run pytest tests for the Django/Ninja backend — service-level tests, API endpoint tests with mocked requests, fixture conventions, and mocking external services (Airbyte, Prefect, warehouses, Redis). Use when writing tests, adding tests, debugging test failures, or asked to test an endpoint or service.

Connect to a client's warehouse (Postgres or BigQuery) from service-layer code using `WarehouseFactory.get_warehouse_client(org_warehouse)`. Use when the code needs to read or write a client's data, execute SQL against their warehouse, inspect schema, or work with `OrgWarehouse`. Never use the legacy `old_client` namespace.

Generate and maintain Docusaurus documentation for Dalgo features. Use when creating new doc pages, updating existing docs, or reviewing documentation for completeness. Provides feature-to-route mapping, doc repo structure, and sidebar conventions.