Skip to content

SKILL.md

Latest commit

 

History

History
267 lines (188 loc) · 10.8 KB

SKILL.md

File metadata and controls

267 lines (188 loc) · 10.8 KB
Raw
name pp-airflow-admin
description Printing Press CLI for Airflow Admin. Focused read-first API surface for inspecting Apache Airflow DAGs, DAG runs, task instances, pools, variables
author Dhilip Subramanian
license Apache-2.0
argument-hint <command> [args] | install cli|mcp
allowed-tools Read Bash
metadata
openclaw
requires install
bins
airflow-admin-pp-cli
kind bins module
go
airflow-admin-pp-cli
github.com/mvanhorn/printing-press-library/library/developer-tools/airflow-admin/cmd/airflow-admin-pp-cli

Airflow Admin — Printing Press CLI

Prerequisites: Install the CLI

This skill drives the airflow-admin-pp-cli binary. You must verify the CLI is installed before invoking any command from this skill. If it is missing, install it first:

  1. Install via the Printing Press installer: 
    npx -y @mvanhorn/printing-press-library install airflow-admin --cli-only
  2. Verify: airflow-admin-pp-cli --version
  3. Ensure $GOPATH/bin (or $HOME/go/bin) is on $PATH.

If the npx install fails (no Node, offline, etc.), fall back to a direct Go install (requires Go 1.26.6 or newer):

go install github.com/mvanhorn/printing-press-library/library/developer-tools/airflow-admin/cmd/airflow-admin-pp-cli@latest

If --version reports "command not found" after install, the install step did not put the binary on $PATH. Do not proceed with skill commands until verification succeeds.

Apache Airflow is a workflow orchestrator used by data teams to schedule and monitor data pipelines. Data engineers usually model each pipeline as a DAG, then watch DAG runs, task instances, retries, pools, and scheduler health to understand whether daily loads, API extracts, dbt jobs, reports, and other data workflows are running correctly.

Use this skill when a user needs an operational readout from Airflow: list DAGs, inspect failed DAG runs, check task instance status, review pools, verify API health, or sync/search Airflow metadata locally. The CLI is read-first and intended for investigation, triage, and reporting. Avoid implying that it deploys DAGs, edits schedules, clears tasks, or mutates Airflow state.

Common Airflow Workflows

Local Airflow auth

For a local Airflow webserver at http://localhost:8080, get a token and store it:

airflow-admin-pp-cli apache-airflow-admin-auth --username airflow --password airflow --json
airflow-admin-pp-cli auth set-token <access_token>
airflow-admin-pp-cli doctor --json

For a remote Airflow environment, set the base URL first:

export AIRFLOW_ADMIN_BASE_URL="https://airflow.example.com"
export AIRFLOW_ADMIN_BEARER_AUTH="<access_token>"

Never ask the user to paste Airflow credentials into a repository, issue, PR, or shared log.

Pipeline health triage

Start broad, then drill down:

airflow-admin-pp-cli monitor --agent
airflow-admin-pp-cli dags list --agent --select dag_id,is_active,is_paused,last_parsed_time
airflow-admin-pp-cli dags dag-runs list <dag_id> --state failed --agent --select dag_id,dag_run_id,state,start_date,end_date
airflow-admin-pp-cli dags dag-runs list-task-instances <dag_id> <dag_run_id> --state failed --agent --select task_id,state,try_number,start_date,end_date

Local search and analysis

When a user wants repeated investigation without paging the Airflow API each time:

airflow-admin-pp-cli sync --resources dags,pools --agent
airflow-admin-pp-cli search "failed" --data-source local --agent --limit 20
airflow-admin-pp-cli analytics count --type dag-runs --agent

Command Reference

apache-airflow-admin-auth — Manage apache airflow admin auth

  • airflow-admin-pp-cli apache-airflow-admin-auth — Authenticate with an Airflow username and password and return a JWT access token.

apache-airflow-admin-version — Manage apache airflow admin version

  • airflow-admin-pp-cli apache-airflow-admin-version — Get Airflow version

connections — Airflow connection metadata.

  • airflow-admin-pp-cli connections — List connections

dags — DAG inventory and metadata.

  • airflow-admin-pp-cli dags get — Get DAG
  • airflow-admin-pp-cli dags list — List DAGs
  • airflow-admin-pp-cli dags dag-runs list — List DAG runs
  • airflow-admin-pp-cli dags dag-runs get — Get DAG run
  • airflow-admin-pp-cli dags dag-runs list-task-instances — List task instances
  • airflow-admin-pp-cli dags dag-runs get-task-instance — Get task instance
  • airflow-admin-pp-cli dags tasks list — List DAG tasks
  • airflow-admin-pp-cli dags details — Get detailed DAG metadata

monitor — Manage monitor

  • airflow-admin-pp-cli monitor — Get health

pools — Airflow pool capacity and occupancy.

  • airflow-admin-pp-cli pools get — Get pool
  • airflow-admin-pp-cli pools list — List pools

variables — Airflow variable metadata.

  • airflow-admin-pp-cli variables — List variables

Finding the right command

When you know what you want to do but not which command does it, ask the CLI directly:

airflow-admin-pp-cli which "<capability in your own words>"

which resolves a natural-language capability query to the best matching command from this CLI's curated feature index. Exit code 0 means at least one match; exit code 2 means no confident match — fall back to --help or use a narrower query.

Auth Setup

Airflow JWT tokens can be created through the API token endpoint when the Airflow environment allows username/password auth:

airflow-admin-pp-cli apache-airflow-admin-auth --username <username> --password <password> --json

Store the returned access_token:

airflow-admin-pp-cli auth set-token YOUR_TOKEN_HERE

Or set AIRFLOW_ADMIN_BEARER_AUTH as an environment variable. Set AIRFLOW_ADMIN_BASE_URL when the Airflow webserver is not http://localhost:8080.

Run airflow-admin-pp-cli doctor to verify setup.

Agent Mode

Add --agent to any command. Expands to: --json --compact --no-input --no-color --yes.

  • Pipeable — JSON on stdout, errors on stderr

  • Filterable — --select keeps a subset of fields. Dotted paths descend into nested structures; arrays traverse element-wise. Critical for keeping context small on verbose APIs:

    airflow-admin-pp-cli apache-airflow-admin-version --agent --select id,name,status

  • Previewable — --dry-run shows the request without sending

  • Offline-friendly — sync/search commands can use the local SQLite store when available

  • Non-interactive — never prompts, every input is a flag

  • Explicit retries — use --idempotent only when an already-existing create should count as success

Response envelope

Commands that read from the local store or the API wrap output in a provenance envelope:

{
  "meta": {"source": "live" | "local", "synced_at": "...", "reason": "..."},
  "results": <data>
}

Parse .results for data and .meta.source to know whether it's live or local. A human-readable N results (live) summary is printed to stderr only when stdout is a terminal AND no machine-format flag (--json, --csv, --compact, --quiet, --plain, --select) is set — piped/agent consumers and explicit-format runs get pure JSON on stdout.

Agent Feedback

When you (or the agent) notice something off about this CLI, record it:

airflow-admin-pp-cli feedback "the --since flag is inclusive but docs say exclusive"
airflow-admin-pp-cli feedback --stdin < notes.txt
airflow-admin-pp-cli feedback list --json --limit 10

Entries are stored locally at ~/.local/share/airflow-admin-pp-cli/feedback.jsonl. They are never POSTed unless AIRFLOW_ADMIN_FEEDBACK_ENDPOINT is set AND either --send is passed or AIRFLOW_ADMIN_FEEDBACK_AUTO_SEND=true. Default behavior is local-only.

Write what surprised you, not a bug report. Short, specific, one line: that is the part that compounds.

Output Delivery

Every command accepts --deliver <sink>. The output goes to the named sink in addition to (or instead of) stdout, so agents can route command results without hand-piping. Three sinks are supported:

Sink Effect
stdout Default; write to stdout only
file:<path> Atomically write output to <path> (tmp + rename)
webhook:<url> POST the output body to the URL (application/json or application/x-ndjson when --compact)

Unknown schemes are refused with a structured error naming the supported set. Webhook failures return non-zero and log the URL + HTTP status on stderr.

Named Profiles

A profile is a saved set of flag values, reused across invocations. Use it when a scheduled agent calls the same command every run with the same configuration - HeyGen's "Beacon" pattern.

airflow-admin-pp-cli profile save briefing --json
airflow-admin-pp-cli --profile briefing apache-airflow-admin-version
airflow-admin-pp-cli profile list --json
airflow-admin-pp-cli profile show briefing
airflow-admin-pp-cli profile delete briefing --yes

Explicit flags always win over profile values; profile values win over defaults. agent-context lists all available profiles under available_profiles so introspecting agents discover them at runtime.

Exit Codes

Code Meaning
0 Success
2 Usage error (wrong arguments)
3 Resource not found
4 Authentication required
5 API error (upstream issue)
7 Rate limited (wait and retry)
10 Config error

Argument Parsing

Parse $ARGUMENTS:

  1. Empty, help, or --help → show airflow-admin-pp-cli --help output
  2. Starts with install → ends with mcp → MCP installation; otherwise → see Prerequisites above
  3. Anything else → Direct Use (execute as CLI command with --agent)

MCP Server Installation

  1. Install the MCP server: 
    go install github.com/mvanhorn/printing-press-library/library/developer-tools/airflow-admin/cmd/airflow-admin-pp-mcp@latest
  2. Register with Claude Code: 
    claude mcp add airflow-admin-pp-mcp -- airflow-admin-pp-mcp
  3. Verify: claude mcp list

Direct Use

  1. Check if installed: which airflow-admin-pp-cli If not found, offer to install (see Prerequisites at the top of this skill).
  2. Match the user query to the best command from the Unique Capabilities and Command Reference above.
  3. Execute with the --agent flag: 
    airflow-admin-pp-cli <command> [subcommand] [args] --agent
  4. If ambiguous, drill into subcommand help: airflow-admin-pp-cli <command> --help.