Tensorlake Skill
A skill that pulls live Tensorlake documentation so agents write sandbox code against real SDK signatures.
- 172 ★ on the repository
- Last commit
- MIT
Extended access
Why this level
- The skill fetches documentation over the network, and the code it prepares uses your Tensorlake API key.
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/tensorlake-skillDetects 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.
- The skill fetches documentation over the network, and the code it prepares uses your Tensorlake API key.
This is third-party code. Review the repository files before installing.
What it does
Tensorlake is a cloud of sandboxes for running agent code. The skill triggers on mentions of Tensorlake and sandboxes: it checks that the SDK is installed and `TENSORLAKE_API_KEY` is set, then loads the `https://docs.tensorlake.ai/llms.txt` index and the relevant documentation pages. If the docs cannot be fetched, the agent falls back to the bundled snapshot in `references/`. Before proposing code it verifies every symbol against live docs or the installed package and never prints API keys.
Who it is for. Developers building agents on Tensorlake sandboxes who are tired of SDK methods the agent invents.
Good fit when
- An agent has to run untrusted Python in an isolated sandbox.
- A sandbox needs to be suspended and resumed later, or forked from a snapshot.
- Parallel sandboxes are launched with `AsyncSandbox` and `asyncio.gather`.
- A port, tunnel or SSH access is opened in a sandbox.
Not a fit when
- The project does not use Tensorlake; the skill is not meant for other clouds.
- The session runs offline and docs cannot be fetched.
- Running code locally without isolation is enough.
Example request
Build a Tensorlake sandbox that runs untrusted Python and exposes port 8000.Limitations
You need your own `TENSORLAKE_API_KEY` from cloud.tensorlake.ai. The skill requests docs.tensorlake.ai on each task; offline only the snapshot in `references/` remains, and it may lag behind the SDK. The skill version is tracked separately from the SDK version. The skill itself runs no code; sandbox runs and their cost follow your Tensorlake plan. There is no keyless mode: access to sandboxes depends on your account plan. The feature list lives in the Tensorlake docs and changes with the SDK.
How to disable. Delete the skill folder from `~/.claude/skills` or your agent's directory. Remove `TENSORLAKE_API_KEY` from your environment separately.
FAQ
Which agents are supported?
Claude Code and Google ADK read `SKILL.md`, OpenAI Codex uses `AGENTS.md`. Through `npx skills add` the skill also installs into Cursor, Cline, GitHub Copilot and Windsurf.
Why does the skill fetch docs on the fly?
Models know the fresh Tensorlake API poorly and invent signatures. Live docs serve as the source of truth; the snapshot in `references/` is the fallback.
How do I set up access?
Get a key at cloud.tensorlake.ai and run `tl login`, or set `TENSORLAKE_API_KEY` in your shell profile or `.env`.
Does the skill cover Python only?
No. It covers the Python SDK (`pip install tensorlake`), the TypeScript SDK (`npm install tensorlake`) and the CLI.
Does the skill start sandboxes itself?
No, it only directs the agent to documentation and checks code against it. Your code creates sandboxes through the SDK under your key.
Discussion
No questions or reviews yet.
Related
SkillsThe official catalog of NVIDIA agent skills: CUDA, accelerated libraries, RAG, robotics and simulation through coding agents
agent-skills (misaka-mikoto-tech)
agent-skills
Three skills by one author for an agent on Windows: safe PowerShell, reading Bilibili videos and working with an open Unity Editor
Skills for writing, debugging, measuring and optimizing TileLang and CUDA programs from source, lowering and profiler evidence
Skill Zoo finds, installs, updates and audits skills for Claude Code, Codex, Cursor and other agents in a window and a CLI.
A skill and docs for building AstrBot plugins: plugin API, events, messages, platform adapters and docs4agent writing rules
A library of 30-plus instruction skills for coding agents: planning, building, testing, hallucination guards and shipping
README in short
A lightweight routing skill: the repository holds `SKILL.md`, `AGENTS.md`, a Claude Code plugin manifest and SDK reference files. CI compares the snapshot with live docs weekly.
README badge
Original SKILL.md
---
name: tensorlake
license: MIT
description: >
Tensorlake SDK — sandboxes, Cloud Volumes, and Git repos for AI agents. Use
when the user mentions tensorlake or sandboxes, or asks about Tensorlake
APIs/docs/capabilities. Also use when building an application, coding agent,
or agentic system that needs a sandbox to run code — e.g., executing
LLM-generated or untrusted code, suspend/resume, snapshots for forking
parallel workers, pre-warmed sandbox pools, custom or OCI images, exposed
ports, egress allowlists, PTY shells, process logs, computer-use / desktop
automation, browser automation (Chrome CDP, Playwright), tunnels for
non-HTTP protocols, async parallel sandboxes, Harbor evals or RL rollouts,
SSH access, remote-dev, or agent integrations (Claude managed agents,
OpenCode, Crabbox, Devin Outposts). Also covers versioned Cloud Volumes and
managed Git repositories mounted into sandboxes, sandbox-native durable
orchestration, and public application endpoints. When this skill applies,
ALWAYS WebFetch https://docs.tensorlake.ai/llms.txt first.
metadata:
author: tensorlake
version: 3.0.1
---
# What can you do with Tensorlake SDK
Tensorlake provides three APIs:
- **Sandbox** — stateful execution environments for AI agents and isolated tool calls, with suspend/resume, snapshots, copy, and pre-warmed pools for persistence between tasks.
- **Orchestration** — sandbox-native durable workflow orchestration for AI agents.
- **Storage** — versioned Cloud Volumes (`tl fs`) and managed Git repositories (`tl git`) that mount into a sandbox as an ordinary directory and outlive it.
Available in **Python**, **TypeScript**, and **CLI**. Use standalone or as infrastructure alongside any LLM provider, agent framework, database, or API.
## Before you start
Verify setup
1. **SDK installed?** If not, install by
**Python:** `pip install tensorlake`
**TypeScript:** `npm install tensorlake`
**CLI:** `curl -fsSL https://tensorlake.ai/install | sh`
2. **API key set?**
For using CLI only, run `tl login`
For using SDKs, get a key at [cloud.tensorlake.ai](https://cloud.tensorlake.ai). and `export TENSORLAKE_API_KEY=your-api-key-here`
## Where to find docs