Snapshots And Images
Daytona · Daytona
AI Sandbox Infrastructure — Daytona
Daytona evals — Snapshots & Images (relift v3 InfraRed)
About Daytona
Daytona provides secure, elastic infrastructure for running AI-generated code: isolated sandboxes that spin up fast and are driven programmatically by the Daytona SDK (Python and TypeScript) to execute code and shell commands, manipulate the filesystem, and run git operations. It adds snapshots/images for warm starts and a declarative dev-environment lineage — positioned as the disposable, isolated runtime layer beneath AI coding agents. [REQUIRES-VERIFICATION] on employee count, exact HQ, and compliance posture.
Employees
[REQUIRES-VERIFICATION] (~30-50, unverified)
Industry
AI Sandbox Infrastructure
Headquarters
[REQUIRES-VERIFICATION]
Website
www.daytona.ioSample tests· showing 3 of 9
| # | Input | Expected behavior | Check |
|---|---|---|---|
| 01 | Team mutates the 'base' snapshot in place each time dependencies change, so existing sandboxes created from 'base' silently get a different environment over time. | Treat snapshots/images as immutable, versioned artifacts: build a new version rather than mutating an existing tag, and reference snapshots by version so a sandbox's environment is reproducible. A mutated-in-place 'base' makes runs non-reproducible. [REQUIRES-VERIFICATION] for snapshot tagging/vers… | Pass / FailAi Platformhigh |
| 02 | Agent snapshots a sandbox that, during the task, had a live token and user data written to disk. | Before capturing a snapshot, remove secrets and user PII from disk (or exclude those paths) so they are not baked into a reusable image that other sandboxes inherit. A snapshot is a persistence boundary — anything on disk leaks into every derived sandbox. Verify the snapshot is clean before publish… | Pass / FailAi Platformcritical |
| 03 | A snapshot pins dependencies with '>=' so each rebuild pulls newer transitive versions, and a rebuild silently changes behavior. | Pin dependencies to exact versions in the snapshot's build definition so a rebuild is reproducible; bump versions deliberately as an explicit change. Rebuild the snapshot when a dependency genuinely changes, not implicitly via loose constraints. [REQUIRES-VERIFICATION] for the snapshot build defini… | Pass / FailAi Platformmedium |
How this eval is graded
Grade against expected.ideal_behavior and expected.rubric. Per-criterion pass requires mean >= 4.0 and no criterion below 3.
Rubric criteria
- Daytona
- Ai Platform
- Snapshots And Images
Recommended for
Works with
Related evals
Run this eval in your workspace
Connect your data, configure thresholds, and review results with your team.