AI Agent Hub
Back to plugins
⚙️

dsh-lines

Workflow Updated 2026.08.25

Run the following command in DeepSeek Harness:

dsh plugin install mario03690/dsh-lines

Paste the following prompt into your AI chat to install this plugin:

Run dsh plugin install mario03690/dsh-lines in the DeepSeek Harness terminal to install the plugin; the full source is available at https://github.com/mario03690/dsh-lines

About this plugin

You have a sequence of tool calls that already works end-to-end, but you need it to fire on a schedule—daily, weekly—without keeping a machine alive just for that. dsh-lines packages that proven call chain into a named pipeline hosted on the ainetcafe cloud scheduler. Close your laptop; the pipeline still triggers on time, each run emits a step-by-step work-order receipt for audit, and downstream steps read upstream outputs via $prev.field so the whole chain stays deterministic with no human in the loop.

Beyond one-shot pipelines, dsh-lines ships watch-style scheduled tasks: declare what to check, how often, and a callback URL, and the cloud handles execution while alerting you only when the result actually changes. The what_can_you_do entry point accepts a plain-language task in any language and returns the exact tool plus a ready-to-run call, eliminating doc-hunting guesswork. All nine tools communicate over MCP; nothing executes locally, there is no filesystem, shell, or process access, and the sole network egress is an outbound HTTPS request to ainetcafe.com.

Who it fits: developers and workflow orchestrators who need cron-grade repetition of multi-step tool chains, want a per-run audit trail, and prefer to be pinged only when something shifts. The free tier requires no signup and covers seven days or sixty runs, whichever comes first; documents and prompts are processed in memory and discarded, with nothing retained server-side.

Use Cases

  • Package a proven multi-step tool-call sequence into a named pipeline that fires on a fixed interval and emits a per-step work-order receipt
  • Run a daily reachability check in the cloud and get notified only when the result actually changes, eliminating noise
  • Use what_can_you_do to describe a task in plain language and immediately get the exact tool plus a ready-to-run call

Best For

  • Developers who need cron-grade repetition of multi-step tool chains without dedicating a machine to run them
  • Platform and SRE engineers who require a per-run audit receipt and care about process traceability
  • Workflow orchestrators who want to be alerted only on genuine state changes and want to cut down on noise