Provides integration guidance and decision-making help for developers who want to add Sandbox0 sandboxing to their AI agents. The skill maps common integration goals to concrete compositions (REPL vs one-shot, volumes for persistence, network restrictions, ports, webhooks) and points to in-repo references and docs for implementation details.
Use when designing or integrating sandboxed execution into an agent: selecting templates or images, choosing persistence strategies, setting network policies, exposing preview servers, handling sandbox events via webhooks, or deciding between managed vs self-hosted Sandbox0 deployments. Ideal for agent maintainers, platform engineers, and SDK users.
agents/ and references/ with detailed integration maps and docs.references/integration-map.md and references/docs-src/ sections for real-world examples and manifest-based navigation.REPL + Volume for persistent coding sessions).Targeted at agent developer workflows and CI/DevOps tooling (supports guidance for CLI/SDK usage, containerized sandboxes, Kubernetes-based deployments).
Sandbox0 Integration is a documentation-only skill that guides AI agent developers through Sandbox0 sandboxing integration via CLI/SDK. No scripts to execute or audit — it's purely reference-based, pointing users to integration-map.md and docs-src/ pages. Clean, no security concerns, but limited to a niche product audience.
Pure documentation/reference skill. No code execution surface. Well-structured progressive disclosure pattern (read map, then specific page). Niche but practical for its target audience.
