Menu
Operate camera streams and realtime inference with model selection, status checks, and logging tools.
Instalar com Codex ou Claude Copie este prompt, cole no Codex, Claude ou outro assistente e deixe que ele revise a página da skill e instale para você.
Baseado na classificação ocupacional SOC
Label video behavior segments from frame-grid images with a defined behavior list, model-ready JSON, resumable outputs, and no_behavior handling for sparse labels.
Use Annolid GUI tools for robust automation across web, PDF, video, and chat controls.
Infer the assay or paradigm from video context, task text, tracked entities, and experimental cues such as social interaction, open field, courtship, resident-intruder, and novel object recognition.
Choose assay-specific features and measurable objectives for behavior analysis, including distances, zones, speed, contact, orientation, and object interaction.
Segment behavior timelines from tracks, pose, contact, speed, and proximity signals into typed intervals with stable labels and rationales.
Keep behavior analysis aligned with assay objectives, controls, reproducibility, and measurable outputs instead of ad hoc summaries.
| name | Camera Realtime Control |
| description | Operate camera streams and realtime inference with model selection, status checks, and logging tools. |
Use these tools for camera/video stream setup and realtime inference automation in Annolid.
gui_list_realtime_models: List available realtime model presets before starting.gui_check_stream_source: Probe camera/stream connectivity before realtime start.gui_start_realtime_stream: Start realtime stream with source/model/transport and optional bot-report settings.gui_get_realtime_status: Verify stream is running and inspect active source/model/viewer.gui_list_realtime_logs: Retrieve detection and bot-event log file paths.gui_list_logs: Inspect top-level Annolid log targets (logs/realtime/runs/label_index/app).gui_open_log_folder: Open a log target in file browser.gui_remove_log_folder: Remove a log target recursively (explicit user request only).gui_stop_realtime_stream: Stop stream cleanly when done.camera_snapshot: Non-GUI camera probe + snapshot saver for chat/email channels.gui_list_realtime_models and choose model.gui_check_stream_source first.gui_start_realtime_stream with explicit source and rtsp_transport (tcp for unstable RTSP).gui_get_realtime_status.gui_list_realtime_logs.gui_stop_realtime_stream after analysis.For "check camera, save snapshot, and email it" requests:
gui_check_stream_source with:
camera_sourcesave_snapshot=trueemail_to=<recipient>email_subject, email_contentoksnapshot_pathsnapshot_opened_on_canvasemail_sent, email_resultok=false or snapshot_path is empty, report probe failure and do not claim email was sent.email_to should be a valid address. When email_to is provided, snapshot capture is required.
When GUI tools are unavailable (for example in email channels):
camera_snapshot with explicit camera_source.snapshot_path as attachment_paths in the email tool.camera_source="0" (or another index).rtsp_transport="tcp" if frames are unstable.rtp://@host:port source and validate receiver/network path first.When starting realtime stream, use:
bot_report_enabledbot_report_interval_secbot_watch_labels (CSV or list)bot_email_reportbot_email_toKeep reporting intervals conservative (>=5s) for high-FPS streams.