workflow-chain

Workflow Chain

Orchestrate multi-step workflows by chaining skills and actions.

Workflow Definition

Workflows are YAML files in ~/.echo-agent/workflows/:

name: morning-routine
trigger: cron "0 8 * * *"
steps:
  - id: get_weather
    skill: weather
    input: {location: "Beijing"}

  - id: get_todos
    skill: reminder
    action: due

  - id: compose
    skill: daily-briefing
    input:
      weather: "{{get_weather.output}}"
      todos: "{{get_todos.output}}"

  - id: deliver
    action: send
    channel: telegram
    content: "{{compose.output}}"

Security

Only run trusted workflow files. The engine runs the commands listed in the JSON workflow definition. Never run workflow files from untrusted sources (downloads, user uploads, model-generated content) without manual review.

Each command is parsed with shlex and executed as a single program with a fixed argument list — shell syntax is not interpreted. ;, &&, |, > and $(...) are passed through as literal arguments, so one step cannot chain into another command. A step that genuinely needs a pipeline or redirection has to ask for a shell explicitly:

{"command": "sh -c 'grep ERROR app.log | wc -l > count.txt'"}

Writing it that way keeps the intent visible in the workflow file instead of every step carrying implicit shell power.

Features

• Variable interpolation: {{step_id.output}} references previous step results
• Conditional steps: when: "{{get_weather.temp}} < 5"
• Error handling: on_error: skip | abort | retry(3)
• Parallel steps: parallel: [step_a, step_b]

Built-in Templates

Research Pipeline

name: research-pipeline
steps:
  - {id: search, skill: web-search, input: {query: "{{topic}}"}}
  - {id: extract, skill: web-extract, input: {urls: "{{search.top_urls}}"}}
  - {id: report, skill: deep-research, input: {content: "{{extract.output}}"}}

Morning Routine

name: morning-routine
trigger: cron "0 8 * * *"
steps:
  - {id: weather, skill: weather, input: {location: "Beijing"}}
  - {id: briefing, skill: daily-briefing}
  - {id: send, action: send, channel: auto, content: "{{briefing.output}}"}