astronomer/agents
27 skills19.1k installs11.1k starsGitHub
Install
npx skills add https://github.com/astronomer/agentsSkills in this repo
1Analyzing DataThe analyzing-data skill is designed for queries data warehouse and answers business questions about data. Handles questions requiring database/warehouse queries including "who uses X", "how many Y", "show me Z",. Data Analysis Answer business questions by querying the data warehouse. All CLI commands below are relative to this skill's directory. Invoke when the user asks about analyzing data or related SKILL.md workflows.1.2kinstalls2AirflowThe airflow skill queries manages and troubleshoots Apache Airflow using the af CLI Covers listing DAGs triggering runs reading task logs diagnosing failures debugging DAG import errors checking connections variables pools and monitoring health Also routes to sub-skills for writing DAGs debugging deploying and migrating Airflow 2 to 3 Use when user mentions Airflow DAG DAG run task log import error parse error broken DAG or asks to trigger a pipeline debug import errors Airflow Operations Use af commands to query manage and troubleshoot Airflow workflows Astro CLI The Astro CLI https www astronomer io docs astro cli overview is the recommended way to run Airflow locally and deploy to production It provides a containerized Airflow environment that works out of the box bash Initialize a new project astro dev init Start local Airflow webserver at http localhost 8080 astro dev start Parse DAGs to catch errors quickly no need to start Airflow astro dev parse Run pytest against your DAGs astro dev pytest Deploy to production astro deploy Full1.2kinstalls3Authoring Dagsauthoring-dags is an official skill from astronomer/agents that walks developers through writing and validating Apache Airflow DAGs using best practices and the af CLI. The skill activates when users create new DAGs, write pipeline code, or ask about DAG patterns and conventions, and it references the companion testing-dags skill for test-debug-fix-retest workflows. A Stop hook reminds developers to test DAGs with testing-dags after authoring. authoring-dags fits data engineers standardizing task dependencies, operators, and project layout so pipelines pass code review before scheduling. It bridges DAG design and the Astronomer CLI commands needed to validate structure during development.991installs4Debugging Dagsdebugging-dags is an Airflow DAG diagnosis skill from astronomer/agents that performs comprehensive failure root cause analysis on data pipelines using the af CLI. The skill triggers on complex debugging requests like diagnose and fix the pipeline, full root cause analysis, or why is this failing and how to prevent it—beyond simple log peeks handled by the airflow entrypoint skill. debugging-dags follows a systematic data-engineering investigation approach to identify root causes and deliver actionable remediation and prevention recommendations. Developers reach for debugging-dags when Airflow DAG failures need structured CLI-driven diagnosis rather than surface-level status checks.984installs5Testing Dagstesting-dags is an Astronomer agents skill for complex Apache Airflow DAG testing that goes beyond a single run command. It handles iterative test-debug-fix cycles when requests include testing a DAG and repairing failures, troubleshooting pipeline issues, or validating multi-step workflows. Simple one-shot test or run requests are delegated to the airflow entrypoint skill, while testing-dags owns repeated validation loops. Commands assume the af CLI on PATH, obtainable via astro otto or uv tool install astro-airflow-mcp. Developers reach for testing-dags when an Airflow DAG fails validation and needs structured debugging with Astro CLI quick checks.953installs6Migrating Airflow 2 To 3migrating-airflow-2-to-3 is an Astronomer agent skill that guides safe migration of Apache Airflow 2.x projects to the Airflow 3.x series. It should load as the first step for any migration-related request involving upgrade paths, compatibility issues, or breaking changes in DAG code. The skill covers DAG files, operators, hooks, and configuration updates, and suggests running ruff check with the AIR rule selector after edits to catch Airflow-specific issues. Developers reach for migrating-airflow-2-to-3 when they detect Airflow 2.x patterns, plan a version upgrade, or need help resolving breaking changes while modernizing scheduled data workflows.942installs7Tracing Upstream Lineagetracing-upstream-lineage is a data-lineage skill from astronomer/agents that helps developers answer "where does this data come from?" by inspecting Airflow DAGs. The workflow starts by identifying whether the target is a table, column, or entire DAG, then searches DAG definitions to find the producing pipeline and upstream data sources. It supports questions about upstream dependencies, data sources, and column-level origins in scheduled Airflow environments. Reach for tracing-upstream-lineage when a warehouse table's provenance is unknown, a column's source DAG must be identified, or an Airflow DAG's read dependencies need mapping before changing pipeline logic.894installs8Managing Astro Local Envmanaging-astro-local-env is an Astronomer agents skill for operating local Apache Airflow environments through the Astro CLI. It supports Docker mode (default, container-based) and Standalone mode (Docker-free, local venv requiring Airflow 3 and uv). Developers use managing-astro-local-env to start, stop, or restart Airflow, tail logs, query the Airflow API, and diagnose environment failures without leaving the agent session. The skill defers new project scaffolding to setting-up-astro-project and points to authoring-dags MCP tools once Airflow is running.887installs9Setting Up Astro Projectsetting-up-astro-project is a DevOps skill for data engineers starting Apache Airflow projects on Astronomer Astro using the Astro CLI. The skill walks through creating a new project, installing Python dependencies, configuring Airflow connections and variables, and understanding the standard directory layout for DAGs, plugins, and settings. It explicitly defers running the local environment to the managing-astro-local-env skill and DAG authoring to authoring-dags, keeping scope on first-time project bootstrap. When teams are not on Astro, the skill points to Apache Airflow Docker Compose for local development and the Helm chart for Kubernetes deployments as open-source alternatives. Reach for setting-up-astro-project when onboarding a pipeline repo, standardizing Astro conventions, or migrating from ad-hoc Airflow installs to managed Astro project structure before writing orchestration code, scheduling production DAG runs, or connecting external data sources.876installs10Profiling Tablesprofiling-tables is an astronomer/agents skill that generates a deep statistical profile of one database table so a new engineer can understand structure, content, and data quality quickly. Step one queries INFORMATION_SCHEMA.COLUMNS for column names, data types, and comments, resolving unqualified table names via INFORMATION_SCHEMA.TABLES. Subsequent steps compute distribution stats, null rates, and quality signals tailored to the target warehouse. Developers invoke profiling-tables when joining a project, auditing an ETL source, or answering data-quality questions before modeling. The skill requires an explicit table name and produces onboarding-ready documentation from live SQL results.874installs11Tracing Downstream Lineagetracing-downstream-lineage answers what breaks if a table, view, or DAG changes by tracing downstream data lineage and impact before edits land in production. The workflow identifies direct consumers by searching DAG source code for SELECT references, using Astronomer CLI commands such as af dags list and af dags source to enumerate and inspect DAG definitions. For tables, the skill guides discovery of downstream DAGs, views, and reports that read the target asset. Developers reach for tracing-downstream-lineage when assessing change risk, planning schema migrations, or validating that a pipeline modification will not break unknown dependents.874installs12Checking Freshnesschecking-freshness is an Astronomer agents skill for fast data currency checks when users ask if data is up to date, when a table was last updated, or whether datasets are stale. For each table, the skill locates timestamp columns such as _loaded_at, _updated_at, created_at, load_date, etl_timestamp, ingestion_time, or business dates like event_date. It fits upstream of analytics queries and agent workflows that assume fresh warehouse tables. Developers reach for checking-freshness before running pipelines, dashboards, or autonomous agents on tables with unknown load schedules.854installs13Annotating Task Lineageannotating-task-lineage is an Astronomer agents skill that walks developers through adding manual lineage metadata to Airflow tasks via `inlets` and `outlets` when operators do not emit OpenLineage events automatically. It references the Apache Airflow OpenLineage provider developer guide for supported patterns and covers Astro deployments where lineage annotations drive downstream catalog visibility. Reach for this skill when a DAG task reads or writes datasets that Marquez, Datakin, or Astro lineage UI must track but the operator has no native extraction. The agent guides dataset URI specification, inlet/outlet placement, and validation against OpenLineage conventions.812installs14Cosmos Dbt Corecosmos-dbt-core is an Astronomer/agents implementation checklist for wiring dbt Core into Airflow via astronomer-cosmos. The skill targets Cosmos 1.11+ and Airflow 3.x, with Appendix A notes for Airflow 2.x import differences. Eight ordered steps cover ProjectConfig, RenderConfig load modes (dbt_manifest, dbt_ls, dbt_ls_file, automatic), ExecutionConfig modes from WATCHER through KUBERNETES, ProfileConfig with SnowflakeUserPasswordProfileMapping or profiles.yml, TestBehavior settings, operator_args, and final DbtDag or DbtTaskGroup assembly. Safety checks enforce secrets via Airflow connections, correct load mode for containerized execution, and BigQuery-only constraints on AIRFLOW_ASYNC. Reach for cosmos-dbt-core when a dbt Core manifest or project path exists and you need Cosmos DAG code—not dbt Fusion, which requires cosmos-dbt-fusion instead.810installs15Airflow Hitlairflow-hitl is an Astronomer agents skill for human-in-the-loop workflows in Apache Airflow 3.1+. It covers ApprovalOperator, HITLOperator, HITLBranchOperator, HITLEntryOperator, and HITLTrigger so DAGs pause until a human responds via the Airflow UI or REST API. HITL operators are deferrable—they release worker slots while waiting for approval, rejection, form input, or human-driven branching. Humans respond from Browse → Required Actions or the task instance Required Actions tab. The skill does not cover AI or LLM calls (see airflow-ai). Use airflow-hitl when pipeline steps need human gates without blocking workers.801installs16Creating Openlineage Extractorscreating-openlineage-extractors is an Astronomer agents skill for implementing custom OpenLineage extractors on Apache Airflow operators that lack built-in lineage support. The skill follows the OpenLineage provider developer guide to handle unsupported or third-party operators, column-level lineage, and extraction logic beyond simple inlets and outlets declarations. Data engineers reach for creating-openlineage-extractors when Marquez, DataHub, or other lineage backends need accurate task-level metadata from custom hooks, sensors, or vendor operators. The workflow covers extractor class structure, operator hook integration, and testing patterns so pipelines emit complete lineage events during DAG execution rather than silent gaps in observability graphs.787installs17Cosmos Dbt Fusioncosmos-dbt-fusion is a configuration reference skill for running dbt Fusion projects through Astronomer Cosmos on Apache Airflow. Fusion support is limited to ExecutionMode.LOCAL with Snowflake or Databricks warehouses, and the skill documents ProfileConfig mappings such as SnowflakeUserPasswordProfileMapping alongside operator_args settings and Airflow 3 compatibility notes. Data engineers reach for cosmos-dbt-fusion when wiring dbt-snowflake or Databricks adapters into Cosmos operators, defining warehouse connections, and aligning DAG tasks with Fusion constraints. It reduces trial-and-error when migrating Cosmos setups from classic dbt Core projects to Fusion-only execution paths.757installs18Warehouse Initwarehouse-init is an Astronomer agents skill that initializes warehouse schema discovery for data teams. Invoked via /astronomer-data:warehouse-init or explicit setup requests, it discovers databases, schemas, tables, and metadata then writes a comprehensive user-editable .astro/warehouse.md file for instant lookups. CLI commands live under the analyzing-data skill's scripts/cli.py directory and require cd into ../analyzing-data/ before execution. Run once per project and refresh when the warehouse schema changes. Developers reach for warehouse-init when analysts and engineers need a searchable schema map tied to repo context instead of ad-hoc INFORMATION_SCHEMA queries. The output becomes the canonical reference agents and humans query during pipeline and modeling work.671installs19Deploying Airflowdeploying-airflow from astronomer/agents teaches reliable deployment of Airflow DAGs and entire projects to production. The skill contrasts Astronomer Astro—managed operations with CLI commands and GitHub integration—for faster CI/CD against open-source paths using Docker Compose for development and the Helm chart for production Kubernetes. Triggers include deploy code, push DAGs, set up CI/CD, deploy to production, and Airflow deployment strategy questions. Developers choose Astro when they want managed ops or self-host with Compose locally and Helm in prod. The skill fits data engineering teams shipping scheduled pipelines who need repeatable promotion from dev clusters to production namespaces without manual tarball uploads.595installs20Troubleshooting Astro Deploymentstroubleshooting-astro-deployments is a Claude Code skill for production incident response on Astronomer Astro. It walks developers through Astro CLI commands such as astro deployment list to locate deployments, inspect logs, analyze scheduler and worker failures, resolve DAG import errors, and adjust deployment environment variables. The skill complements managing-astro-deployments for lifecycle management and managing-astro-local-env for local reproduction. Developers reach for it when Airflow jobs fail after deploy, workers disappear from the cluster, or environment-specific runtime errors block scheduled DAG runs. It assumes familiarity with Astro-managed Airflow rather than self-hosted clusters.445installs21Airflow Pluginsairflow-plugins is an Astronomer agent skill for constructing Apache Airflow 3.1+ plugins that extend the scheduler and web UI with custom operators, hooks, sensors, listeners, macros, middleware, and operator extra links. The skill walks through AirflowPlugin registration patterns, embedding FastAPI-backed endpoints and React apps via fastapi_apps, external_views, and react_apps, serving static assets, and calling the Airflow REST API from inside a plugin. Developers reach for airflow-plugins when they need a custom nav entry, bespoke UI page, or API middleware inside Airflow rather than a standalone service. Triggers include mentions of AirflowPlugin, plugin registration, embedding a web app in the Airflow UI, or building pipeline automation that must live natively inside Airflow 3.1+.441installs22Managing Astro Deploymentsmanaging-astro-deployments is an agent skill from astronomer/agents that encodes production deployment workflows for Astronomer Astro using the Astro CLI. The SKILL.md walks through OAuth login with astro login, workspace switching, listing and inspecting deployments, and lifecycle operations to create, update, or delete production environments with executor, HA, and resource settings. It documents four deploy modes—full, DAG-only, image-only, and force—and shows how to promote DAGs from staging to production, rotate deployment API tokens for CI/CD, and enable fast DAG-only iteration. Developers reach for managing-astro-deployments when Airflow pipelines are ready to ship on Astro and they need repeatable CLI steps instead of guessing deployment flags. The skill pairs with managing-astro-local-env for local dev and troubleshooting-astro-deployments for production incidents.422installs23Blueprintblueprint is an Astronomer agent skill for the airflow-blueprint PyPI package that composes Apache Airflow DAGs from YAML using reusable Python task-group templates with Pydantic validation. It requires Python 3.10+, Airflow 2.5+, and Blueprint 0.3.0+, enabling validated configurations and no-code DAG authoring for teams that want repeatable pipeline patterns. Developers reach for blueprint when creating blueprint templates, composing DAGs from YAML, validating Airflow configs, or letting non-engineers author DAGs from approved building blocks. The skill follows ordered implementation steps and prefers the simplest configuration that meets pipeline needs, drawing from the astronomer/blueprint open-source repository.369installs24Dag Factorydag-factory helps teams scaffold and evolve Airflow DAGs through agent-friendly patterns from Astronomer, turning repeatable pipeline definitions into generated orchestration code for ETL, analytics, and operational jobs.239installs25Migrating Ai Sdk To Common AiThe migrating-ai-sdk-to-common-ai skill from astronomer/agents guides Claude through refactoring agent code from a proprietary AI SDK to Astronomer's Common AI layer. It maps API differences, updates imports and configs, and preserves tool calls and streaming so migrations stay safe and incremental.210installs26Delegating To Ottodelegating-to-otto from astronomer/agents teaches parent agents when and how to route data-orchestration requests to Otto, Astronomer's specialist agent. It supports building multi-agent systems that manage Airflow DAGs and Astro Cloud operations reliably.186installs27Airflow AdapterThis skill documents an adapter pattern that makes code compatible with both Airflow 2.x (/api/v1) and 3.x (/api/v2) APIs. It is used when working with the adapters, adding version-detecting API methods, or handling field and endpoint differences between the two Airflow versions. The version is auto-detected at startup and routed to the matching adapter.2installs