dstack
A unified orchestration layer for heterogeneous AI compute. It standardizes how to manage compute and run training and inference on GPU clouds, Kubernetes, VMs, or bare-metal clusters.
No outside reports on file
dstack touches cloud credentials, Kubernetes access, and job configurations; use it only with trusted infrastructure, limit its permissions, audit what it can reach, and monitor resource usage closely.
Based on public adoption and transparency signals as of Oct 21, 2026. Our opinion, not a code audit or security guarantee.
What it can access
- Local filesNo local folder access notedLOW
- Command executionRuns shell or system commandsREVIEW
- Network reachConnects to outside service APINOTED
- CredentialsRequires API key or tokenREVIEW
- Package sourceGitHub (dstackai/dstack)PUBLIC
Inferred from this listing’s public package details and description. Check the project’s source before installing.
Install
No npm or PyPI package is listed. Install from source: github.com/dstackai/dstack ↗
About Dstack
dstack is a tool that helps manage AI computing resources across different types of hardware and cloud platforms. It lets teams run machine learning training and inference jobs on GPUs, Kubernetes clusters, VMs, or physical servers using a single unified system instead of learning many different tools.
Who it’s for
Machine learning engineers and teams who work with multiple cloud providers or on-premise hardware and want a consistent way to manage their AI workloads. It's designed for organizations running large-scale model training or inference.
Alternatives
- Kubernetes Job/Pod scheduling: Native Kubernetes for simpler workloads and existing clusters. - Ray Cluster management: Distributed computing framework popular for ML workloads. - Slurm: Traditional HPC job scheduler used in research and on-premise environments. - Cloud provider ML services: AWS SageMaker, Azure ML, or Google Vertex AI for single-cloud solutions.
Setup
Check these steps against the project’s own instructions before running anything.
Follow the project's README at https://github.com/dstackai/dstack (opens external site) for installation and configuration instructions, as setup varies by your infrastructure type (cloud, Kubernetes, bare metal, etc.).
Why this grade
Signal-based grade (3/7 points): well starred on GitHub, public source code. This reflects adoption and transparency, not a code review.
Popularity and upkeep
- GitHub stars
- 2.3k
- Listing last refreshed
- 2026-10-29
These are public popularity signals, not a maintenance check. Look at the repository’s commit history and open issues.
Security reports
No reports found for this package in our research (last checked 2026-10-29).
Risks to watch for
- Cloud credential access: dstack needs credentials to your cloud accounts (AWS, Azure, GCP) to provision and manage resources; a compromise could expose your computing infrastructure. - Kubernetes permissions: If used with Kubernetes, it requires cluster admin or high-level permissions that could be abused to access other workloads or data. - Configuration injection: Malicious YAML or config files could trick the tool into running unexpected code or exposing secrets stored in your infrastructure. - Resource hijacking: An attacker with access to dstack could spin up expensive GPU instances or redirect your jobs to steal training data. - Logs and outputs: Job logs and model outputs may contain sensitive information like training data, API keys, or business logic if not carefully controlled.
How to lower the risk
Least-privilege permissions
- Grant dstack only the minimum cloud permissions needed (separate roles for provisioning, monitoring, cleanup).
- Use role-based access control in Kubernetes; avoid cluster-admin permissions.
Network isolation
- Run dstack in a restricted network segment away from sensitive production systems.
- Use VPCs or private Kubernetes namespaces to isolate compute jobs.
Monitoring
- Log all dstack API calls and resource creation events.
- Track GPU and compute spending for sudden spikes indicating unauthorized use.
- Audit job configurations and outputs for leaked secrets or unexpected data.
Red flags to watch for
- Unexpected cloud credentials or environment variables in job configs.
- Large numbers of GPU instances spinning up without your request.
- Job logs containing API keys, passwords, or training data in plain text.
Questions
How do I install Dstack?
Dstack does not publish a standalone npm or PyPI package. Clone or follow the setup instructions in its source repository at github.com/dstackai/dstack.
Are there security reports about Dstack?
No reports found for this package in our research (last checked 2026-10-29). Dstack currently holds Grade C (Moderate risk signals) based on public adoption and transparency signals. Key risk areas to watch include cloud credential access.
Is Dstack actively maintained?
Dstack has 2.3k GitHub stars. Our directory data was last refreshed on 2026-10-29. Check recent commit activity, release tags, and open issues at github.com/dstackai/dstack before pinning it in production.
What are alternatives to Dstack?
Notable alternatives to Dstack include Kubernetes Job/Pod scheduling (Native Kubernetes for simpler workloads and existing clusters. - **Ray Cluster management**: Distributed computing framework popular for ML workloads. - **Slurm**: Traditional HPC job scheduler used in research and on-premise environments. - **Cloud provider ML services**: AWS SageMaker, Azure ML, or Google Vertex AI for single-cloud solutions). You can also compare Dstack side by side with other skills in the Related tools section on this page.
README badge and team notes
Maintain or use Dstack? Show its live grade in your GitHub README, docs site, or run a check from the terminal before you install.
GitHub README Markdown:
[](https://agentarmorhq.com/skills/dstackai-dstack)HTML embed:
<a href="https://agentarmorhq.com/skills/dstackai-dstack"><img src="https://agentarmorhq.com/api/badge/dstackai-dstack" alt="AgentArmorHQ Risk Grade for Dstack" height="20" /></a>Terminal check:
curl -fsSL https://agentarmorhq.com/api/cli | node --input-type=module - check "dstackai-dstack"Share with your security lead
Need sign-off to use dstack? Copy this note for Slack, a ticket or a pull request.
### AgentArmorHQ Security Review Memo: dstack
- **Listing URL:** https://agentarmorhq.com/skills/dstackai-dstack
- **Type:** AI agent skill
- **Risk Grade:** Grade C (Moderate risk signals)
- **Outside Security Advisories on File:** 0
- **Public Adoption:** 2.3k GitHub stars
#### Permission & Blast-Radius Signals
- **Local files:** No local folder access noted
- **Command execution:** Runs shell or system commands
- **Network reach:** Connects to outside service API
- **Credentials:** Requires API key or token
- **Package source:** GitHub (dstackai/dstack)
#### Recommended Hardened Setup
- **Source Repo:** https://github.com/dstackai/dstack
- **Controls:** Pin an exact version in `.mcp.json`, scope access to a single workspace folder, and use a least-privilege read-only token where supported.
_Based on public adoption and transparency signals as of Oct 21, 2026. Our opinion, not a code audit or security guarantee._Key facts
- Type
- AI agent skill
- Publisher
- dstackai
- Source code
- github.com/dstackai/dstack
- Est. context load
- ~72 tokens idle · ~450 tokens when active (on-demand skill)
- Popularity
- 2.3k GitHub stars
- Last updated
- 2026-10-29
- Last graded
- 2026-10-21
Community
Discussions for Dstack
Ask questions or share configuration tips. Governed by our Community Rules & Terms.
No posts match this filter yet.
Be the first to ask a question or share a working tip.
Check before you install, in Claude
Your assistant answers Block, Review or No known issues for any package before it installs it.
Get this check inside Claude