blaxel-ai

Add or update a Blaxel Sandbox Hub image/template under hub/. Use when adding a new hub image, sandbox image, runtime image, template.json, Dockerfile, hidden/internal image, or workflow-dispatch build option in the blaxel-ai/sandbox repository.

Deploy a Playwright sandbox (chromium or firefox) on Blaxel and run e2e tests against it. Use to validate that playwright hub images work end-to-end (browser connection, page navigation, DOM interaction) after changes.

Poll CI status on a branch until the "Build and Push Sandbox" workflow completes. Use after pushing to develop or any branch where you need to confirm the build succeeded before proceeding.

Watch CI until the "Build and Push Sandbox" workflow completes. For dev deployments, watches the develop branch. For prod deployments, watches the latest tag. Use after merging into develop or pushing a release tag.

Deploy the app to the dev environment. Use after a PR is marked as reviewable (ready for review) so reviewers can check out the changes running in dev.

Start the local sandbox-api development environment with hot-reload. Use when developing or testing changes to the sandbox-api Go code locally.

Run the full end-to-end test suite against a custom sandbox image deployed locally or on Blaxel. Use before merging significant changes to validate the complete sandbox-api binary, not just unit/integration tests.

Run the sandbox-api integration tests against a live API instance. Use after making changes to sandbox-api to verify all endpoints still work correctly.

Use when migrating sandbox code from Daytona (@daytonaio/sdk or daytona_sdk) to Blaxel (@blaxel/core or blaxel). Covers API mapping for sandbox creation, process execution, filesystem operations, and authentication in both TypeScript and Python. Reach for this skill when you encounter Daytona SDK imports or need to replace Daytona sandbox calls with Blaxel equivalents.

Use when migrating sandbox code from E2B (@e2b/code-interpreter or e2b) to Blaxel (@blaxel/core or blaxel). Covers API mapping for sandbox creation, command execution, filesystem operations, pause/resume, and authentication in both TypeScript and Python. Reach for this skill when you encounter E2B SDK imports or need to replace E2B sandbox calls with Blaxel equivalents.

Use when migrating sandbox code from Modal (modal SDK) to Blaxel (@blaxel/core or blaxel). Covers API mapping for sandbox creation, command execution, filesystem operations, image building, snapshot/resume, tunnels, and authentication in both TypeScript and Python. Reach for this skill when you encounter Modal SDK imports or need to replace Modal sandbox calls with Blaxel equivalents.

Build the CLI in development mode. Use when the user says "build", "build dev", "compile", "rebuild", or when you need to test local changes to the CLI binary. Also use this after writing code changes to verify they compile correctly, or when the user wants to try out their changes locally.

Lint all changed Go files, fix every error, and confirm the build passes. Use when the user says "fix lint", "run lint", "lint", or when pre-commit hooks fail. Also use this proactively before committing code or when you notice linter warnings in build output.

Manage Blaxel resources from the command line using the bl CLI. Deploy agents, sandboxes, jobs, and MCP servers. Also installs the Blaxel CLI if not present.

Use when creating cloud sandboxes (microVMs) to run code, start dev servers, and generate live preview URLs. Also covers deploying AI agents, MCP servers, batch jobs, and Agent Drives (shared filesystems) on Blaxel's serverless infrastructure. Reach for this skill when you need isolated compute environments, real-time app previews, shared file storage across sandboxes, or to deploy agentic workloads.

Compare sandbox "time to first interaction" between the current git branch and main in the Blaxel sdk-typescript repo. Runs the cold-call benchmark (create sandbox, first call, delete) on both branches, then reports a side-by-side table. Use when the user says "compare sandbox bench", "compare ttfi", "bench PR vs main", "run sandbox bench compare", "time to first interaction", or provides a PR/branch and asks to benchmark it against main. Builds @blaxel/core before each run. Caps total sandbox usage to stay under 50.