Provision per-user containers (Docker/Firecracker) with Node, Python, ROS, FPGA, and RTOS toolchains pre-installed. Stream terminals + file edits directly into the browser UI.
- In progress
Container images
Base + ROS/micro-ROS + FPGA variants published to registry.
- Planned
Browser IDE shell
Expose VS Code Web + multiplexed bash/PowerShell sessions per learner.
- Planned
State sync
Persist workspaces via Supabase buckets + JSONL run logs.
Extend /api/ai endpoints so copilots can inspect repos, propose diffs, run tests, and summarize failures with audit trails.
Depends on: Phase 1
- Planned
Workspace context feed
Secure channel streaming file trees + terminal output to the tutor.
- Planned
Command broker
API for tutor-issued workspace.exec requests with human approval gates.
- Planned
Action ledger
Log every AI intervention to JSONL for review + payouts.
Ship a lightweight agent that exposes WebUSB/WebSerial, JTAG, oscilloscope, and logic-analyzer feeds back to the workspace UI and AI tutor.
Depends on: Phase 1
- Planned
Device discovery
Detect Arduino, ESP32, STM32, FPGA boards, and RTOS targets via USB.
- Planned
Flashing + telemetry
Provide safe firmware flashing with automatic log capture + rollback.
- Planned
Virtual instruments
Render scope/logic data in-browser when physical gear unavailable.
Mission manifests define repo templates, BOMs, checkpoints, and payout triggers so every build log ties into the ledger.
Depends on: Phase 1 → Phase 2 → Phase 3
- Planned
Manifest schema
Describe hardware kits, datasets, and test harness per mission.
- Planned
Test adapters
Support Jest, rostest, Zephyr, MicroPython, Verilator harnesses.
- Planned
Ledger hooks
Emit mission-complete events so payouts + ops dashboards stay in sync.