Sandbox automation console

openai-agents-sandbox-demo

Sandboxed automation runner for testing agent patches, manifests, streamed logs, and credential-gated execution.

SandboxAutomationOpenAI Agents
Healthblocked
Agents1attached workers
Runs176% success
Artifacts2retained outputs
24h cost$0.00estimated spend
Next runWaiting for OPENAI_API_KEYschedule

Integration plan

Treat this as a guarded automation surface, separate from normal dashboard agents.

Source/Users/advithgopinath/dev/PERSONAL/openai-agents-sandbox-demo
Repositoryhttps://github.com/AdvithGopinath/openai-agents-sandbox-demo
Deploymentsandbox://local

Risk register

Keep blockers visible before scheduling more automations.

RiskCredential missing for true e2e
RiskNeeds artifact retention

Agents

Persistent workers scoped to this project surface.

All agents
Sandbox Runnerblocked

Runs manifest-bound coding automations in isolated workspaces.

No broad filesystem access; only manifest-staged files.

Installed templates

Reusable library items configured for this project surface.

Library
No installed templatesplanned

Install a library template to track project-specific reusable agents, automations, tools, or Codex profiles.

Browse library

Runs

Latest executions for this surface.

Calculator fixture patchValidated compile and manifest shape, then stopped because OPENAI_API_KEY was missing.
failed7m 02s

Activity

Project-scoped event stream.

Sandbox run blockedThe dry run stopped before e2e execution because OPENAI_API_KEY is missing.

Trend

Recent run volume and health trend.