# dltHub Toolkits Cheatsheet

> Agentic Workflows

## Complete agentic workflows for every phase of data engineering

Not autocomplete, not a chatbot on a dashboard. A guided sequence of skills, commands, rules, and MCP - with guardrails agents can't skip. Maintained by dltHub,  controlling the infrastructure agents and pipelines operate on.

## REST API Pipeline

Connect to any API and load data automatically

7 skills | 1 rule | MCP enabled

### /find-source

Find a dlt source for a given API or data provider. Use when the user asks about a source, wants to find a connector, or asks to implement a pipeline for a specific data source.

Steps:
1. Classify request (sql_database, rest_api, filesystem)
2. Search verified sources with dlt init --list-sources
3. Search dlthub catalog via MCP
4. Web search for API docs
5. Report source with endpoints and auth details

Example prompts:
- Use find-source to load data from Attio CRM into DuckDB

### /create-rest-api-pipeline

Create a dlt REST API pipeline. Use for the rest_api core source, or any generic REST/HTTP API source.

Steps:
1. Snapshot current folder
2. Run dlt init rest_api <destination>
3. Research API docs
4. Write pipeline code
5. Configure secrets
6. Run with add_limit(1)
7. Report loaded data

Example prompts:
- Create a REST API pipeline for Attio companies and load into DuckDB

### /new-endpoint

Add a new REST API endpoint/resource to an existing dlt pipeline.

Steps:
1. Read existing pipeline code
2. Research new endpoint
3. Add resource to RESTAPIConfig
4. Update __main__ with_resources
5. Run and validate

Example prompts:
- Add deals endpoint to the existing Attio pipeline

### /adjust-endpoint

Adjust a working dlt pipeline for production — remove dev limits, verify pagination, configure incremental loading.

Steps:
1. Check for .add_limit() calls
2. Verify explicit paginator config
3. Configure incremental loading
4. Remove dev_mode=True
5. Run full load with INFO logging

Example prompts:
- Remove dev limits and add pagination to the Attio pipeline

### /validate-data

Validate schema and data after a successful dlt pipeline load. Check row counts, schema, data types.

Steps:
1. Export schema as mermaid
2. View tables and row counts
3. Sample rows
4. Fix data types
5. Iterate until correct

Example prompts:
- Validate the loaded Attio data looks correct

### /view-data

Query, explore, or view data loaded by a dlt pipeline. Covers dlt dataset API, ibis, and ReadableRelation.

Steps:
1. Attach to pipeline
2. Use pipeline.dataset() API
3. Query with ibis
4. Show results as DataFrame

Example prompts:
- Show me the top companies from the Attio pipeline

### /debug-pipeline

Debug and inspect a dlt pipeline after running it. Inspect traces, load packages, schema, and diagnose errors.

Steps:
1. Set log_level to INFO
2. Run pipeline
3. Inspect trace
4. Check load packages
5. Fix root cause

Example prompts:
- Debug the Attio pipeline — it failed with a 401 error

## dltHub Runtime

Deploy to production with one command

4 skills | 2 rules

### /setup-runtime

Verify dlt workspace is ready for dltHub Runtime. Use when deploying for the first time.

Steps:
1. Verify pyproject.toml
2. Check .dlt/.workspace
3. Verify dlt[hub] dependency
4. Run dlt runtime login
5. Confirm ready

Example prompts:
- Set up my workspace for dltHub Runtime

### /prepare-deployment

Prepare production credentials and destinations. Set up profile-scoped secrets and production destinations.

Steps:
1. Verify .dlt/ config
2. Create prod.secrets.toml
3. Configure destination
4. Test credentials

Example prompts:
- Set up production credentials for the Attio pipeline

### /deploy-workspace

Deploy dlt pipelines to dltHub Runtime. Assumes workspace is verified and credentials are set.

Steps:
1. Remove dev_mode and limits
2. Verify write_disposition
3. Pin dlt version
4. Deploy + launch
5. Verify logs

Example prompts:
- Deploy the Attio pipeline to production

### /debug-deployment

Debug a failed or misbehaving dltHub Runtime deployment. Check job status and logs.

Steps:
1. Check job status
2. View logs
3. Inspect run details
4. Fix and redeploy

Example prompts:
- The Attio pipeline failed on runtime, help me debug it

## Data Exploration

Explore data locally, build notebooks, ship Marimo dashboards

2 skills | 1 rule

### /explore-data

Explore dlt pipeline data locally. Connect, profile tables, plan charts with ibis + altair, and write an analysis plan.

Steps:
1. Attach to pipeline
2. Profile tables
3. Detect intent
4. Plan chart (ibis + altair)
5. Write analysis_plan.md

Example prompts:
- Explore my Attio data — show companies by employee range

### /build-notebook

Assemble a marimo notebook from analysis_plan.md. Reads chart specs, generates Python file, validates, and launches.

Steps:
1. Read analysis_plan.md
2. Parse chart specs
3. Generate dashboard.py
4. Validate notebook
5. Launch with marimo run

Example prompts:
- Build the dashboard notebook from the analysis plan

## Transformations

Transform raw data into a Canonical Data Model

4 skills | 1 rule | MCP enabled

### /annotate-sources

Annotate dlt pipeline sources for transformation. Map data sources to canonical business concepts.

Steps:
1. Confirm pipeline names via MCP
2. Extract schemas
3. Classify columns
4. Map to business concepts
5. Identify natural keys
6. Write taxonomy.json

Example prompts:
- Annotate my Attio pipeline sources for a sales analytics CDM

### /create-ontology

Build a business entity graph from annotated sources and taxonomy.

Steps:
1. Read taxonomy + annotations
2. Build entity list
3. Define relationships
4. Choose master sources
5. Write ontology.json

Example prompts:
- Build the entity graph for the sales analytics CDM

### /generate-cdm

Generate a Canonical Data Model using Kimball dimensional modeling. Star schema from your ontology.

Steps:
1. Read ontology
2. Classify as fact/dimension
3. Design star schema
4. Generate CDM.dbml
5. Validate

Example prompts:
- Generate the star schema CDM for sales analytics

### /create-transformation

Write @dlt.hub.transformation functions that map source tables to CDM entities using ibis.

Steps:
1. Read taxonomy + CDM
2. Map source to CDM
3. Write transformation
4. Test output
5. Validate against schema

Example prompts:
- Write transformations to map Attio companies to the dim_company entity

## Foundation

Cross-toolkit rules, secrets management, and agent routing

3 skills | 1 rule | 1 command | MCP enabled

### /improve-skills

Improve existing skills based on the current session. Capture debugging patterns, doc references, workflow improvements.

Steps:
1. Scan session for learnings
2. Identify skills to update
3. Read current SKILL.md
4. Add lean updates
5. Append to right section

Example prompts:
- Improve the skills based on what we learned in this session

### /setup-secrets

Safely manage dlt secrets in *.secrets.toml. Uses MCP tools to never expose raw values.

Steps:
1. Identify required credentials
2. Check existing with secrets_list
3. Update with secrets_update_fragment
4. Verify with secrets_view_redacted

Example prompts:
- Set up the Attio API token in my secrets

### /toolkit-dispatch

Route users to the right toolkit and skill. Use when the user asks "what can you do" or "where do I start".

Steps:
1. Discover toolkits via MCP
2. Get skill details
3. Match intent to skill
4. Suggest install if needed

Example prompts:
- What can I do with dlt?