Security audit
Security checks for vulnerabilities and agentic risk
The skill performs the advertised HRV video analysis, but it silently creates or reuses identity state, contacts external services, and stores authentication tokens for sensitive biometric/health-adjacent data.
Review before installing. This skill may upload facial video or video URLs to lifeemergence.com/open.lifeemergence.com services, automatically associate reports with an internal identity, create local workspace data, and persist service tokens for future history retrieval. Install only if users have consented to biometric/health-adjacent processing and you are comfortable with cloud report history and local credential persistence.
This finding combines multiple undeclared behaviors: remote API communication, identity/user creation and token handling, local file and workspace writes, and database interaction, while the description frames the skill as a simple facial-video health monitor. In a biometric/health context, hidden identity and storage behaviors materially increase privacy, compliance, and data-exfiltration risk.
This finding combines multiple undeclared behaviors: remote API communication, identity/user creation and token handling, local file and workspace writes, and database interaction, while the description frames the skill as a simple facial-video health monitor. In a biometric/health context, hidden identity and storage behaviors materially increase privacy, compliance, and data-exfiltration risk.
This finding combines multiple undeclared behaviors: remote API communication, identity/user creation and token handling, local file and workspace writes, and database interaction, while the description frames the skill as a simple facial-video health monitor. In a biometric/health context, hidden identity and storage behaviors materially increase privacy, compliance, and data-exfiltration risk.
This finding combines multiple undeclared behaviors: remote API communication, identity/user creation and token handling, local file and workspace writes, and database interaction, while the description frames the skill as a simple facial-video health monitor. In a biometric/health context, hidden identity and storage behaviors materially increase privacy, compliance, and data-exfiltration risk.
This finding combines multiple undeclared behaviors: remote API communication, identity/user creation and token handling, local file and workspace writes, and database interaction, while the description frames the skill as a simple facial-video health monitor. In a biometric/health context, hidden identity and storage behaviors materially increase privacy, compliance, and data-exfiltration risk.
This finding combines multiple undeclared behaviors: remote API communication, identity/user creation and token handling, local file and workspace writes, and database interaction, while the description frames the skill as a simple facial-video health monitor. In a biometric/health context, hidden identity and storage behaviors materially increase privacy, compliance, and data-exfiltration risk.
This finding combines multiple undeclared behaviors: remote API communication, identity/user creation and token handling, local file and workspace writes, and database interaction, while the description frames the skill as a simple facial-video health monitor. In a biometric/health context, hidden identity and storage behaviors materially increase privacy, compliance, and data-exfiltration risk.
This finding combines multiple undeclared behaviors: remote API communication, identity/user creation and token handling, local file and workspace writes, and database interaction, while the description frames the skill as a simple facial-video health monitor. In a biometric/health context, hidden identity and storage behaviors materially increase privacy, compliance, and data-exfiltration risk.
This finding combines multiple undeclared behaviors: remote API communication, identity/user creation and token handling, local file and workspace writes, and database interaction, while the description frames the skill as a simple facial-video health monitor. In a biometric/health context, hidden identity and storage behaviors materially increase privacy, compliance, and data-exfiltration risk.
This finding combines multiple undeclared behaviors: remote API communication, identity/user creation and token handling, local file and workspace writes, and database interaction, while the description frames the skill as a simple facial-video health monitor. In a biometric/health context, hidden identity and storage behaviors materially increase privacy, compliance, and data-exfiltration risk.
This finding combines multiple undeclared behaviors: remote API communication, identity/user creation and token handling, local file and workspace writes, and database interaction, while the description frames the skill as a simple facial-video health monitor. In a biometric/health context, hidden identity and storage behaviors materially increase privacy, compliance, and data-exfiltration risk.
YARA rule matched a hack tool or exploit indicator (offensive tools, reconnaissance, privilege escalation, or exploit frameworks).
---
name: "smyx-facial-hrv-trend-monitoring-analysis"
description: "Using everyday cameras (laptop camera, smartphone front camera, smart mirror), the system records 30-60 seconds of facial video and uses remote photoplethysmography (rPPG) to extract subtle color variations from facial skin micro-circulation, from which it computes heart-rate-variability (HRV) metrics including SDNN (standard deviation of all normal sinus RR intervals) and RMSSD (root mean square of successive RR-interval. | 通过日常摄像头（如电脑摄像头、手机前置摄像头）拍摄面部视频（30-60秒），利用光电容积描记技术（远程光电容积�
Persisting token and open_token fields is especially risky because these values often grant authentication or API access. In a health-monitoring skill with no clear authentication-management purpose, storing such secrets locally can enable account takeover, lateral movement, or long-term unauthorized access if the SQLite database is read by another skill or attacker.
The utility reads a workspace key file, derives a default open-id, and creates/persists local database user records when no identity is supplied. That behavior is unrelated to HRV analysis and creates silent identity state that can later be used for remote API actions, enabling tracking, impersonation-by-default, or unauthorized account linkage.
This shared utility performs remote API access, token handling, and account bootstrap logic that is unrelated to the declared local facial HRV analysis purpose. In the context of a camera-based health-monitoring skill, hidden networked identity and token workflows materially increase data-exposure and unauthorized service-use risk, especially because requests are auto-issued without clear user consent at the utility layer.
The code automatically logs in or creates remote accounts using derived local identity state, then stores resulting tokens for future use. For an HRV trend-monitoring skill, this is unjustified privilege expansion: it can create external identities and bind local users to remote services without transparent authorization, enabling account misuse, unexpected data sharing, and persistent tracking.
The skill advertises and operationally depends on shell, file, network, environment, and local persistence capabilities, but the manifest does not declare any explicit tool scope or permission boundaries. In an agent setting, this increases the chance of over-broad execution and makes it harder for reviewers and policy systems to constrain what the skill is allowed to access.
The history-report trigger phrases are broad enough that ordinary user language could unintentionally invoke cloud retrieval of sensitive historical health reports. Because the skill deals with biometric/health-adjacent data, accidental triggering increases privacy exposure and the chance of over-collection.
The skill states that uploaded files are automatically saved locally but does not provide a clear user-facing warning about storage, retention, or security of those files. Since the inputs are facial videos, silent local saving creates a meaningful biometric privacy risk.
The cloud history-query workflow instructs automatic remote API access but does not clearly warn users that sensitive report data will be sent to or retrieved from an external service. For health-related report history, lack of explicit disclosure undermines informed consent and may violate privacy expectations.
The documentation describes API key authentication and report export endpoints for highly sensitive health data, but provides no guidance on privacy controls, authorization scoping, transport security, audit logging, or data minimization. In a biometric/health-monitoring context, this omission increases the risk that implementers expose medical-like reports through weakly scoped keys or inadequately protected export functionality.
The user-facing argparse description and help strings are presented only in Chinese, which imposes a specific language on all users without any opt-in or locale selection. This matches the policy category for language or locale constraints that are not optional or explicitly justified.
The script silently resolves an internal user identity via OpenIdUtil.resolve_current_open_id(...) and exposes a hidden --list mode that retrieves analysis history through skill.get_output_analysis_list(open_id=...), neither of which is necessary for one-off facial HRV processing. In a health-related skill handling sensitive biometric data, hidden identity binding and history enumeration increase the risk of unauthorized access to prior analyses, privacy violations, and cross-user data exposure if identity resolution or authorization is weak.
Large whitespace padding was detected (a block of blank lines or a long run of spaces). This can push injected instructions below or to the right of the visible area so a human reviewer never sees them while the agent still reads them. Manual review of the hidden content is recommended.
result_json = JsonUtil.parse(result_json_pure_text, result_json_pure_text)
result_json_common_ai_response = result_json.get("commonAiResponse") if isinstance(result_json,
dict) else result_json
if result_json_common_ai_response:
result_json = result_json_common_ai_response
Large whitespace padding was detected (a block of blank lines or a long run of spaces). This can push injected instructions below or to the right of the visible area so a human reviewer never sees them while the agent still reads them. Manual review of the hidden content is recommended.
result_json = JsonUtil.parse(result_json_pure_text, result_json_pure_text)
result_json_common_ai_response = result_json.get("commonAiResponse") if isinstance(result_json,
dict) else result_json
if result_json_common_ai_response:
result_json = result_json_common_ai_response
Detected: suspicious.install_untrusted_source