Airflow Read-Only Skill for Openclaw

A secure, read-only interface to inspect Apache Airflow DAGs, runs, and task instances via the Stable REST API.

kansodata
v0.1.0
Apr 9, 2026
1
835
0

Install & Download

1. ClawHub CLI

The fastest way to install a skill directly from the registry.

npx clawhub@latest install kansodata-airflow-readonly

2. Manual Installation

Copy the skill folder to one of these locations

Global
~/.openclaw/skills/
Workspace
<project>/skills/

Priority: Workspace > Local > Bundled

3. Prompt Installation

Copy this prompt to OpenClaw to install it automatically.

Help me install kansodata-airflow-readonly using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).

Prefer to download?

Get the raw skill files in a ZIP archive.

What is Airflow Read-Only Skill?

The Airflow Read-Only skill empowers AI agents to query Apache Airflow environments through the Stable REST API v2 without the risk of modifying production state. Designed for operational transparency, it allows users to safely audit orchestration pipelines and diagnose execution issues directly within the ecosystem of Openclaw Skills. This skill strictly prohibits mutations, ensuring that your DAG configurations and execution schedules remain untouched during inspection.

By providing a controlled read-only layer, this skill acts as a bridge for developers who need real-time data from their Airflow instance to inform decision-making or generate reports. It is a vital tool for maintaining high availability and visibility into complex data workflows while adhering to strict security protocols.

Airflow Read-Only Skill Use Cases

  • Auditing and listing available DAGs across an Airflow instance.
  • Investigating recent DAG run statuses to identify failures or delays.
  • Drilling down into specific task instances to check durations and execution states.
  • Generating automated reports on pipeline health using real-time orchestration data.

How Airflow Read-Only Skill Works

  1. The agent discovers existing workflows by calling the list_dags tool to retrieve active configurations.
  2. Once a specific DAG is identified, the agent retrieves its execution history via the list_dag_runs tool using the exact dag_id.
  3. For a selected execution, the agent fetches granular details of all task instances using the list_task_instances tool with the dag_run_id.
  4. The resulting data is presented in a compact, actionable format for diagnostics or further analysis.

Airflow Read-Only Skill Setup

To enable this capability within Openclaw Skills, ensure the following steps are completed:

  1. Install the required plugin:
npm install @kansodata/openclaw-airflow-plugin
  1. Configure your environment with a valid Airflow host URL.
  2. Provide valid authentication credentials, such as a Bearer Token, to access the /api/v2 endpoint.
  3. Verify that the plugin is active in your Openclaw configuration.

Airflow Read-Only Skill Data Schema & Taxonomy

Object Tool Mapping Description
DAG airflow.list_dags Returns a list of all DAGs including IDs and operational status.
DAG Run airflow.list_dag_runs Lists executions for a specific dag_id, including state and timing.
Task Instance airflow.list_task_instances Provides details for tasks within a run, including duration and state.

All metadata returned adheres to the Apache Airflow Stable REST API schema, ensuring compatibility with standard monitoring tools.

Airflow Read-Only Skill Advanced Features

  • Context-aware chaining: Seamlessly transition from DAG discovery to task-level diagnostics within a single session.
  • Strict read-only enforcement: Guaranteed prevention of accidental triggers, pauses, or configuration changes.
  • Exact identifier handling: Ensures data integrity by using precise dag_id and dag_run_id references without normalization artifacts.
  • Multi-agent compatibility: Designed to work natively with other Openclaw Skills for comprehensive multi-agent diagnostic workflows.

SKILL.md


Loading

Related Openclaw Skills

METADATA

Github Stars: 0
forks: 0

Featured*