InspireSkill skill
A CLI and skill for the Inspire compute platform: an agent creates notebooks, training jobs and services without web clicks.
- 563 ★ on the repository
- Last commit
- MIT
Extended access
Why this level
- Commands occupy paid compute and change images and jobs on the platform.
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/inspireskillDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- Commands occupy paid compute and change images and jobs on the platform.
This is third-party code. Review the repository files before installing.
What it does
InspireSkill turns the web interface of the Inspire platform (Qizhi, SII) into an `inspire` command and ships a skill for it. An agent picks a workspace, project and resources, starts a Notebook, runs GPU, HPC and Ray jobs, deploys Serving and reads events and logs. The `SKILL.md` points the agent to `inspire --help` as the source of truth for flags, while platform details sit in separate references. The installer places the skill into detected clients: Codex, Claude Code, Cursor, OpenCode and others.
Who it is for. Students and researchers with access to the Inspire platform who want to hand resource and job routine to an agent.
Good fit when
- A GPU notebook must be ordered and code synced without a dozen clicks.
- A training job fails and the agent has to collect events and logs.
- You need to know how many free eight-card nodes exist now.
- A model is deployed as a service and QPS and latency are watched.
Not a fit when
- You have no account on the Inspire platform: the tool works only with it.
- Jobs run on another cloud.
- You are not ready to give an agent access to resources your project pays for.
Example request
List the workspaces available to me, find a free eight-GPU node and describe how to start training on it.Limitations
The tool is built on the platform's internal web requests, so changes on the Inspire side can break some commands until an update ships. Documentation is in Chinese. The installer downloads and runs a script from the repository and installs a PyPI package, so read it before running. Commands can occupy expensive GPU resources and change images and models, so check the project quota before a bulk launch.
How to disable. Run `inspire uninstall` to remove the CLI and installed skill files. Data in `~/.inspire` stays; wipe it with `--purge` and `--purge-runtime`.
FAQ
Which clients are supported?
Codex, Claude Code, Cursor, OpenCode, Kimi Code, Qoder, Antigravity, OpenClaw, Pi and others. The target is set with `--harness`.
What does `--no-cli` do?
Installs only `SKILL.md` and the references without the `inspire` command.
How do I update the skill?
With `inspire update --skill-only`. Without the flag the CLI updates too.
What is needed to install?
`bash`, `curl`, `tar`, Python 3.10 or newer, and `uv` or `pipx`.
How do I watch a service after launch?
`serving metrics` shows GPU, CPU and memory load, `serving api-metrics` shows QPS, success rate and latency.
Discussion
No questions or reviews yet.
Related
SkillsThe official set of skills and extensions for agents working on distributed Aspire applications.
A skill that teaches a coding agent to use Dormice sandboxes: acquire by key, run commands, move files and set lifecycle.
Agent Skill Manager
agent-skill-manager
A universal skill manager for coding agents: finds and installs skills, scanning them for shell calls, network access and credential leaks first
An SSH skill for Codex and Claude Code: remote commands, file transfer and tunnels with one JSON result and replay protection.
Cloud.ru Skills for AI Agents
Cloud.ru Skills для AI-агентов
Official Cloud.ru skills: account setup, Foundation Models, ML Inference, virtual machines, RAG and Evolution AI Agents
Official AMD skills teach an agent to set up ROCm, serve LLMs on Instinct and EPYC, profile GPUs and quantize models.
README in short
The README sets out the project goal, installation for macOS, Linux and Windows, updates and the CLI feature list.