Security audit
Security checks for vulnerabilities and agentic risk
The skill is a cloud-connected driver health video analyzer, but it silently creates or reuses an internal identity and stores authentication data while handling sensitive face and health-related records.
Review this before installing. Use it only if you are comfortable sending driver face video or video URLs to the configured lifeemergence.com APIs, creating or reusing an internal identity, and keeping identity/token data in the workspace data directory. Prefer explicit consent from affected drivers and avoid broad history queries unless the operator is authorized to view those reports.
The skill reportedly performs authenticated external API communication, silent identity creation/login, and local workspace file management while claiming to be a visual health monitor. Hidden authentication and account lifecycle behaviors are especially sensitive in a biometric-health context because they can silently link user identity, uploaded facial media, and historical reports.
The skill reportedly performs authenticated external API communication, silent identity creation/login, and local workspace file management while claiming to be a visual health monitor. Hidden authentication and account lifecycle behaviors are especially sensitive in a biometric-health context because they can silently link user identity, uploaded facial media, and historical reports.
The skill reportedly performs authenticated external API communication, silent identity creation/login, and local workspace file management while claiming to be a visual health monitor. Hidden authentication and account lifecycle behaviors are especially sensitive in a biometric-health context because they can silently link user identity, uploaded facial media, and historical reports.
The skill reportedly performs authenticated external API communication, silent identity creation/login, and local workspace file management while claiming to be a visual health monitor. Hidden authentication and account lifecycle behaviors are especially sensitive in a biometric-health context because they can silently link user identity, uploaded facial media, and historical reports.
The skill reportedly performs authenticated external API communication, silent identity creation/login, and local workspace file management while claiming to be a visual health monitor. Hidden authentication and account lifecycle behaviors are especially sensitive in a biometric-health context because they can silently link user identity, uploaded facial media, and historical reports.
The skill reportedly performs authenticated external API communication, silent identity creation/login, and local workspace file management while claiming to be a visual health monitor. Hidden authentication and account lifecycle behaviors are especially sensitive in a biometric-health context because they can silently link user identity, uploaded facial media, and historical reports.
The skill reportedly performs authenticated external API communication, silent identity creation/login, and local workspace file management while claiming to be a visual health monitor. Hidden authentication and account lifecycle behaviors are especially sensitive in a biometric-health context because they can silently link user identity, uploaded facial media, and historical reports.
The skill reportedly performs authenticated external API communication, silent identity creation/login, and local workspace file management while claiming to be a visual health monitor. Hidden authentication and account lifecycle behaviors are especially sensitive in a biometric-health context because they can silently link user identity, uploaded facial media, and historical reports.
The skill reportedly performs authenticated external API communication, silent identity creation/login, and local workspace file management while claiming to be a visual health monitor. Hidden authentication and account lifecycle behaviors are especially sensitive in a biometric-health context because they can silently link user identity, uploaded facial media, and historical reports.
The skill reportedly performs authenticated external API communication, silent identity creation/login, and local workspace file management while claiming to be a visual health monitor. Hidden authentication and account lifecycle behaviors are especially sensitive in a biometric-health context because they can silently link user identity, uploaded facial media, and historical reports.
The skill reportedly performs authenticated external API communication, silent identity creation/login, and local workspace file management while claiming to be a visual health monitor. Hidden authentication and account lifecycle behaviors are especially sensitive in a biometric-health context because they can silently link user identity, uploaded facial media, and historical reports.
The skill reportedly performs authenticated external API communication, silent identity creation/login, and local workspace file management while claiming to be a visual health monitor. Hidden authentication and account lifecycle behaviors are especially sensitive in a biometric-health context because they can silently link user identity, uploaded facial media, and historical reports.
The skill reportedly performs authenticated external API communication, silent identity creation/login, and local workspace file management while claiming to be a visual health monitor. Hidden authentication and account lifecycle behaviors are especially sensitive in a biometric-health context because they can silently link user identity, uploaded facial media, and historical reports.
YARA rule matched a hack tool or exploit indicator (offensive tools, reconnaissance, privilege escalation, or exploit frameworks).
---
name: "smyx-driver-flushing-sweat-detection-analysis"
description: "Using an in-cabin DMS camera, the system analyzes the driver's facial video in real time, detecting skin color variation (flush index, derived from red-channel ratio in RGB or skin-color models) and sweat-droplet / reflective area (via image texture and reflection features). | 通过车载DMS摄像头实时分析驾驶员面部视频，检测面部肤色变化（潮红指数，通过RGB色空间中的红色分量比例或肤色模型）以及汗珠/反光面积（通过图像纹理和反射特征）。当潮红指数显著升高（可能提示血压升高、发热或情绪激动�
The file contains generic user-account persistence logic that is unrelated to the declared driver flushing/sweat video analysis purpose. In a safety/health-monitoring skill, hidden account management increases data collection scope and attack surface, especially because it supports persistent identity records and lookup/update behavior without a clear need tied to the manifest.
The User model stores tokens, open_token, email, realname, birthday, sex, and age, which are highly sensitive and unjustified for a driver flushing/sweat detection skill. Collecting and retaining identity and authentication data in a health-adjacent monitoring context materially increases privacy harm, account compromise impact, and regulatory exposure if the local database is accessed.
The skill claims to perform in-cabin video analysis and health alerts, but this utility also resolves identities from local workspace state, reads a secret-like file, creates default user identities, and persists them. That hidden account lifecycle behavior is outside the declared purpose and can silently bind the user or workspace to backend identities without informed consent, increasing privacy and supply-chain risk.
This HTTP client does more than simple API transport: it performs token recovery, auto-login/registration, local token persistence, and injects identity metadata into outbound requests. For a driver monitoring skill, these hidden remote account and token-management capabilities are unrelated to the stated function and create unnecessary opportunities for data exfiltration, unauthorized account linkage, and backend abuse.
The skill advertises and instructs use of capabilities including shell execution, network access, file read/write, and environment use, but does not declare any tool scope or permission boundaries. This creates an avoidable trust gap where an agent may invoke powerful operations without explicit least-privilege constraints, increasing the chance of unintended data access, command execution, or network exfiltration.
The manifest frames the skill as real-time in-cabin analysis, while the documentation adds cloud history-querying and report-link retrieval functions. This scope drift matters because it introduces extra data access and disclosure pathways for sensitive driver-health records that are not obvious from the primary description.
The default trigger is broad enough that the skill may activate on any driver-face video analysis request, causing automatic handling of sensitive biometric media without strong user intent confirmation. Overbroad auto-invocation increases the chance of unintended uploads, storage, or execution of a higher-risk workflow.
The history-report trigger phrases are broad and automatically invoke cloud lookups, which can expose or retrieve sensitive health-related records with minimal user specificity. In a fleet/driver-health setting, this creates elevated privacy risk because historical biometric-event data may be surfaced too easily.
Automatic creation or reuse of internal user identities is a significant undeclared behavior for a skill whose stated purpose is image/video analysis. In this context, identity binding can silently connect sensitive biometric-health events across sessions and reports, increasing privacy and access-control risk.
The CLI exposes behavior beyond the stated video-analysis function by supporting a hidden identity-backed listing mode via --list and suppressed --open-id handling. This creates an undisclosed capability to enumerate or retrieve prior analysis records tied to a user identity, which can leak sensitive health-related data and violates least privilege and transparency expectations.
The code silently resolves an internal user identity with OpenIdUtil.resolve_current_open_id() even though identity handling is not necessary for local or URL-based video analysis. Hidden identity resolution increases the risk of unauthorized access to user-linked data, especially in a health-monitoring context where outputs may reveal sensitive biometric or medical inferences.
Detected: suspicious.install_untrusted_source