mirror of
https://github.com/runpod/runpod-plugins-official.git
synced 2026-09-14 18:44:38 +08:00
Tim Pietrusky
140f53f402
Per request: evals should run against real infra, no mocks. Reworks the dev-loop eval from description-grade to a live run graded on observed runtime results, and commits a self-contained buggy fixture project. - add flash/evals/fixtures/dev-loop/main.py: an Endpoint with an intentional module-level VOL referenced in the handler (NameError on the remote worker under flash dev; flash deploy masks it) - rewrite dev-loop-iteration.eval.md: Setup (RUNPOD_API_KEY, real worker, cost), observed-result assertions (background dev server, real port from log, file-namespaced /main/predict route, verbatim NameError from worker logs, hot-reload fix, real HTTP 200, undeploy), and Cleanup
Runpod Agent Skills
Skills for AI agents to manage GPU workloads on Runpod.
Available Skills
flash
Complete knowledge of the runpod-flash framework - SDK, CLI, architecture, deployment, and codebase. Use when working with runpod-flash code, writing @remote functions, configuring resources, debugging deployments, or understanding the framework internals.
runpodctl
Manage GPU pods, serverless endpoints, templates, volumes, and models.
Installation
npx skills add runpod/skills
Works with Claude Code, Cursor, GitHub Copilot, Windsurf, Cline, and 17+ other AI agents.
Setup
runpodctl doctor
Usage
Ask your AI agent:
- "Create a pod with an RTX 4090"
- "List my pods"
- "What GPUs are available?"
- "Show my account balance"
- "Deploy a serverless endpoint"
URLs
Pod URLs
Access exposed ports on your pod:
https://<pod-id>-<port>.proxy.runpod.net
Example: https://abc123xyz-8888.proxy.runpod.net
Serverless URLs
https://api.runpod.ai/v2/<endpoint-id>/run # Async request
https://api.runpod.ai/v2/<endpoint-id>/runsync # Sync request
https://api.runpod.ai/v2/<endpoint-id>/health # Health check
https://api.runpod.ai/v2/<endpoint-id>/status/<job-id> # Job status
Structure
flash/
└── SKILL.md
runpodctl/
└── SKILL.md
License
Apache-2.0