kami-suspicious-personDetect unregistered faces loitering in sensitive areas. Supports one OR many RTSP cameras concurrently in a single process (shared ONNX models, per-camera tr...
Install via ClawdBot CLI:
clawdbot install 13681882136/kami-suspicious-personGrade Fair — based on market validation, documentation quality, package completeness, maintenance status, and authenticity signals.
Calls external URL not in known-safe list
https://github.com/deepinsight/insightface/releases/download/v0.7/buffalo_l.zipUses known external API (expected, informational)
open.feishu.cnAudited May 20, 2026 · audit v1.0
Generated Oct 6, 2026
A small retail chain uses the skill on RTSP cameras covering stockrooms and back offices to flag unknown faces lingering after closing. Owners receive JSON alarms via the heartbeat inbox and can review saved snapshots of strangers who stay beyond the loiter threshold.
Property managers monitor entrance cameras to detect unregistered visitors loitering near a rental unit beyond the allowed guest count. The face database holds tenants' pre-approved faces, and any unmatched face that lingers triggers an alert so hosts can verify before escalating.
IT teams deploy the skill on cameras covering restricted server rooms to detect anyone not in the authorized face database who lingers near racks. Since detection runs purely on CPU with ONNX, it runs on existing edge hardware without needing a GPU server.
Site managers watch gates and equipment yards for unregistered faces that remain past the loiter threshold, helping prevent theft and unauthorized access. Multiple RTSP cameras share a single ONNX model process, keeping resource usage low on modest site hardware.
HOA or condominium managers set up the skill across common-area cameras to spot unknown persons loitering near lobbies, mailrooms, or parking structures. Persistent alarm JSON output integrates with their existing monitoring dashboard to notify security staff in real time.
A provider hosts the skill on lightweight edge boxes for small businesses and charges a monthly fee per camera. They manage the face database, tune thresholds, and deliver alerts through a hosted dashboard.
Vendors sell a pre-configured mini-PC or gateway with the skill preinstalled and an admin UI for face enrollment. Customers keep all video and embeddings on-premise, appealing to privacy-conscious enterprises.
Consultancies integrate the skill's JSON alarm output into existing VMS, access control, or IoT platforms for clients with unique loitering rules. They customize thresholds, build face databases, and wire alerts into Slack, SMS, or physical alarms.
💬 Integration Tip
Because the skill emits one-line JSON alarms to stdout and an inbox file, wrap it with a log forwarder or webhook bridge to route events into your existing VMS, Slack, or SMS alerting pipeline. Pre-build the face_db.pkl cache and pin it as a read-only volume so multiple camera processes share a consistent, fast-loading face database.
Scored Oct 6, 2026
Real-time search engine supporting web search, vertical domain search, parallel batch search, and URL content extraction.
Manage Feishu (Lark) calendars by listing, searching, checking schedules, syncing events, and marking tasks with automated date extraction.
cad reference tool
Search, install, and create OpenClaw skills using intelligent matching across built-in, local, and GitHub skill repositories.
Use when building CLI tools, implementing argument parsing, or adding interactive prompts. Invoke for CLI design, argument parsing, interactive prompts, progress indicators, shell completions.
Publish local skills to ClawHub with customizable version, name, path, and optional changelog using command line or Python API.