Intelligent Outdoor Care Monitoring & Analysis Tool | 户外看护智能监测分析工具

Detects targets such as people, vehicles, non-motorized vehicles, and pets within target areas; supports batch image analysis, suitable for outdoor surveilla...

Before installing or running this skill, consider the following: - Data exfiltration: The skill sends uploaded images/videos to remote APIs (API base URLs are configured in skills/smyx_common). If you plan to analyze sensitive camera footage, verify the remote endpoints (check skills/smyx_common/scripts/config.yaml and config-prod/test files) and only use endpoints you trust. - Declared vs actual behavior: The SKILL.md forbids reading local memory, yet the code contains YAML config loaders and a local SQLite DAO that will create/read files in the workspace. Decide whether local persistence is acceptable for your environment. - Unrelated bundled functionality: The package includes a separate face_analysis subskill (TCM face diagnosis) and a large common library. If you only want simple outdoor object detection, request or use a trimmed package that excludes unrelated code to reduce attack surface and privacy concerns (face analysis may enable sensitive biometrics processing). - Environment variables and config: The code reads environment variables (OPENCLAW_SENDER_OPEN_ID, OPENCLAW_WORKSPACE, FEISHU_OPEN_ID) and config files for API keys. The registry metadata lists none of these; do not provide credentials without confirming where they are used. Prefer running in a sandbox and inspect skills/smyx_common/scripts/config.yaml before supplying secrets. - Local file writes: The skill will write config files and a local SQLite DB under a derived workspace path. If you cannot tolerate local persistence, do not run it or set OPENCLAW_WORKSPACE to a safe sandboxed directory and monitor created files. - Test in isolation: Run the skill in an isolated environment (network-limited VM or container) first, monitor outbound network calls, and confirm the API endpoints and behavior match what you expect. - Ask the author for provenance: Because the package source is 'unknown', request the maintainer's homepage, formal API docs, or a minimal distribution that only contains the outdoor monitoring code and exact endpoint URLs. If you cannot verify the origin or endpoints, treat the package as untrusted. What would increase confidence: developer provenance and a trimmed code bundle, clear list of external endpoints and required env vars in the registry metadata, and confirmation that local DB usage will not store or read historical reports (to match the SKILL.md assertions).

Type: OpenClaw Skill Name: smyx-outdoor-monitoring-analysis Version: 1.0.0 The skill bundle provides outdoor monitoring and image analysis by interfacing with a remote cloud service (lifeemergence.com). It utilizes a shared utility layer (smyx_common) that handles user authentication and session management via a local SQLite database (smyx-common-claw.db). While the SKILL.md contains high-priority 'forced memory rules' that instruct the AI agent to ignore local memory files in favor of cloud API data, and the code includes the ability to invoke the openclaw CLI via subprocess, these appear to be functional requirements for maintaining data synchronization across the Smyx ecosystem rather than malicious intent.