Loopify

Turn a manual workflow into a loop.

Describe a recurring job. The Loopify skill drafts the loop.yaml, install guide, README, check script, and catalog entry. It stays draft until a human reviews it and a clean run proves it.

Copy the skillGitHub raw skill

Workflow converted into a loopA manual checklist becomes a bounded trigger, act, verify, and gate loop.check feedswrite notesposttriggeractverifygate
Loopify turns a manual workflow into a draft loop bundle.

What you get

Give your agent the workflow, not a blank page

Loopify starts with Q&A intake, summarizes what is already known, asks only for missing details one question at a time, then parses the routine into trigger, context, act, verify, and escalate. Risky actions become approval gates.

Fetch https://loopmaster-ai.pages.dev/skills/loopify/SKILL.md and follow it.
Interview me about the job I want to automate — one question at a time,
only what you can't infer from this conversation.
Do not schedule, publish, spend, request credential values, or make
destructive changes until I approve.

What Loopify produces

A draft loop bundle is useful because the moving parts are explicit. Your agent gets the manifest and install guide. You get the readable summary and gates.

  • loop.yaml for the loop's job description.
  • README.md for the human explanation.
  • AGENT-INSTALL.md for concrete setup steps.
  • verify.sh for deterministic checks.
  • A draft catalog entry for later publication.

Draft means draft: nothing is scheduled or published until you approve it.

Five-beat loop ringTrigger, context, act, verify, and escalate arranged in a loop.triggercontextactverifyescalateloop
Five beats make a loop.

Worked example: Monday feeds

The checked-in example converts “Every Monday I check three feeds and write a summary” into a watchdog loop. It does not publish by itself. The post step is gated for human approval, and the draft verifier checks the manifest plus sample report shape.

Read the skill page

Human rendering of the Loopify authoring workflow.

Fetch raw SKILL.md

Stable agent-readable skill URL: https://loopmaster-ai.pages.dev/skills/loopify/SKILL.md

Safety gates

Default-deny permissions

The generated manifest starts with the smallest allowed action set.

Approval for risk

Publishing, deleting, overwriting, or spending must be marked as approval-required.

No secrets in manifests

Credentials are declared by scope only. Values stay outside transcripts and files.

Draft until proven

Loopify can draft a loop. Only a human plus a clean run can call it verified.