recamera-intellisense_skill

Overview

This skill registers and manages reCamera devices, configures AI detection models/rules/schedules, monitors and clears detection events, and captures snapshots or video on demand. It drives local Python CLI scripts that communicate with reCamera HTTP APIs using JSON I/O. The skill is designed for automated onboarding, detection automation, event polling, and manual capture tasks.

How this skill works

All operations call the bundled Python CLI scripts under scripts/ with a single JSON object argument. Device credentials are stored in ~/.recamera/devices.json and referenced by device_name or passed inline. Detection configuration, event polling, image fetches, and capture commands return JSON on success; errors appear on stderr and include one actionable fix suggestion. For long-running automations, the skill checkpoints start_unix_ms for incremental event polling.

When to use it

• Onboard or rotate reCamera device credentials
• Configure object-detection models, rules, or schedules locally
• Continuously poll for detection events and fetch event images when needed
• Run manual one-shot image captures or scheduled capture automation
• Clear event storage or troubleshoot detection behavior

Best practices

• Always pass a complete JSON object to CLI scripts; avoid interactive prompts
• Use device_name (preferred) or a single inline device object — never both
• Store per-camera tokens in ~/.recamera/devices.json and set file permissions to 600
• Poll get_detection_events every 1–10 seconds and checkpoint start_unix_ms for incremental reads
• Prefer reading event metadata first; fetch images only when required to save bandwidth and I/O
• Run scripts from the repository base or append ./scripts to sys.path when importing in Python

Example use cases

• Add a new camera and verify with list_devices before automation
• Map a label name to index via get_detection_models_info, then set detection rules for that label
• Run a loop that polls get_detection_events, checkpoints timestamps, and fetches images for matching events
• Trigger capture_image to grab an on-demand snapshot saved to a local path for evidence or debugging
• Clear_detection_events to reset event storage before a testing session