1. Habit
  2. Skills
  3. Runpod MCP skill
runpod · 12K installs

Runpod MCP skill
Runs by itself, every Monday morning.

Weekly Runpod inventory, nothing touched.

Install the Runpod MCP skill and make it a habit: follow https://gethabit.dev/skills/runpod/runpod-mcp.md

  1. 1Copy
  2. 2Paste into Claude Code, Codex or Cursor
  3. 3It runs by itself

12K installs46 starsApache-2.0 licenceSource on GitHub ↗

Runpod MCP

by runpod · 12K installs

The skill

The runpod mcp skill lets an agent manage Runpod pods, serverless endpoints, templates, volumes and billing through the Runpod MCP server, and connect that server when it is missing.

Only when you ask.

The habit

Weekly Runpod inventory, nothing touched.

↻ Runs by itselfevery Monday morning

How the habit works.
You don't have to.

Three questionsWhen does it start? How does it check its work? What does it remember?

01

It starts by itself

every Monday morning. Runs in Claude Code routine.

  • ↻ every Monday morning
  • 0 8 * * 1

02

It checks its work

live read checks. If a check fails, it stops.

  • ✓/mcp shows runpod as Connected
  • ✓a list-endpoints call returns without error
  • ✓every fact in the report names the read behind it
  • ✓the table lists each resource with its name and id

03

It remembers

weekly inventory log. The next run reads it first: resource ids from the last run.

loops/runpod-inventory-loop/runs.md

One run, step by step.
Then it stops.

Stop ruleStops after one account audit per run, with no retries.

  1. 01

    Check connection

    The agent runs /mcp and a list-endpoints call, and falls back to runpodctl if the tools are missing.

  2. 02

    Read last run

    It reads the run log to know which ids it saw before.

  3. 03

    Route the task

    It follows the lifecycle-crud journey for the inventory and cost-audit for billing.

  4. 04

    Read live state

    It confirms every resource and its status with a fresh list or get call.

  5. 05

    Audit only

    For resources this run did not create, it lists ids and the actions you could take, and changes nothing.

  6. 06

    Write the report

    It closes with one table of names, ids and states, each tied to its read.

  7. 07

    Log the run

    It appends the ids and any change since last week to the run log.

What it may do.
Nothing more.

Safe by defaultIt writes only in its own folder and asks you before anything goes out.

Is it for you?
Use it when it fits.

Quick checkRead both lists. If the left one sounds like you, the Runpod MCP skill helps.

Use it when

  • ✓ You manage Runpod pods
  • ✓ You deploy a serverless endpoint
  • ✓ You audit your Runpod account

Skip it when

  • – You need file transfer or SSH
  • – You build Docker images locally

Questions

Runpod MCP skill, answered.

Can the loop delete idle pods?

No. The skill lets the agent change only what it created or what you named. The loop reports, you decide.

What if a permission error comes back?

The skill treats it as out of scope. The loop reports it and does not retry with another key.

How do I install the Runpod MCP skill?

Copy the prompt at the top into your agent. It installs the skill from runpod/runpod-plugins-official and sets up the habit. Or run npx skills add runpod/runpod-plugins-official --skill runpod-mcp.

Is the Runpod MCP skill free?

Yes. It is published under Apache-2.0. Habit is free too.

Runpod MCP. Now as a habit.

Install the Runpod MCP skill and make it a habit: follow https://gethabit.dev/skills/runpod/runpod-mcp.md

  • Free
  • No account
  • Undo in one line

All runpod skills →More ops skills →

Works in

  • Claude Code
  • Codex
  • Cursor
  • ChatGPT
  • Claude app

Made by runpod, published under Apache-2.0. Habit links to the original and never changes it. Is this your skill?

Copying didn't work here. Select the line and copy it by hand.