Skip to content
Beskid Beskid

Beskid

Jump to a Beskid service

Beskid

Jump to a Beskid service

Operate Beskid Learn

Beskid Learn provides interactive learning and runs real CLI checks against temporary learner source.

FieldVerified value
PurposeInteractive learning with analyze, parse, tree, and run checks.
AudienceLearners use lessons. Service operators verify the CLI-backed runtime.
Public boundaryhttps://learn.beskid-lang.org.
Local boundarypnpm run dev --port 4173 starts the frontend. The production container uses port 80.
AuthenticationThe platform Compose contract accepts auth-hub and Learn session settings. Do not assume sign-in when these settings are absent.
Persistent stateLearner checks use a temporary workspace. The production Compose contract declares no durable Learn volume.
Container imageghcr.io/cyber-nomad-collective/beskid-learn.
Health checkGET /api/health on the service port.
Deployment ownerThe root platform delivery publishes and promotes the Learn image.
Secret sourceCurrent required operation has no Learn OpenBao path. Optional auth values come from deployment configuration.
MonitoringContainer health and an API check show whether the web and CLI paths work.
RecoveryRevert to the previous verified image when either the health check or lesson smoke check fails.

Use the interactive learning service for a verified lesson. This safe procedure requires an explicit, known BESKID_BINARY for a local full loop. The service can fall back to cargo run, but do not use that fallback in this procedure.

  1. Verify the service contract at https://learn.beskid-lang.org/api/health.
  2. Record the health response.
  3. Open https://learn.beskid-lang.org.
  4. Select one seeded exercise.
  5. Run the exercise check.
  6. Inspect the real diagnostic or successful result.

The /api/health request succeeds. The service writes the submitted source to a temporary workspace, invokes the selected command, and returns its result.

If the page works but a check fails, verify the deployed CLI and runtime-kit evidence. If all checks fail after a release, restore the prior image digest and redeploy. Do not preserve learner source from the temporary workspace as service state.

Learners should Use Beskid Learn. Service operators should check all service health endpoints.