Back to skill

Security audit

Living-Alone Sleep Rhythm Anomaly Analysis | 独居者作息规律异常分析

Security checks across malware telemetry and agentic risk

Overview

The skill appears to perform the advertised cloud-based video analysis, but it silently manages user identity, stores access tokens locally, and can query sensitive home-monitoring history without clear user control.

Review before installing. Use this only with explicit consent from the monitored person and only if you are comfortable sending overnight in-home video, derived reports, and account-linked history to the configured lifeemergence.com services. Ask the publisher for retention, deletion, report-sharing, and account/token storage details, and avoid untrusted remote video URLs.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Behavioral ASTexec() Call, eval() Call, Dynamic Import
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
Findings (18)

Dynamic attribute access via getattr()

Low
Category
Dangerous Code Execution
Content
if filters:
                for key, value in filters.items():
                    query = query.filter(getattr(self.__model__, key) == value)

            if offset:
                query = query.offset(offset)
Confidence
77% confidence
Finding
query = query.filter(getattr(self.__model__, key) == value)

Dynamic attribute access via getattr()

Low
Category
Dangerous Code Execution
Content
if filters:
                for key, value in filters.items():
                    query = query.filter(getattr(self.__model__, key) == value)

            return query.scalar()
        finally:
Confidence
77% confidence
Finding
query = query.filter(getattr(self.__model__, key) == value)

Lp3

Medium
Category
MCP Least Privilege
Confidence
94% confidence
Finding
The skill exposes capabilities for environment access, file read/write, shell execution, and network access without declaring permissions or clearly constraining their use. In a skill handling highly sensitive bedroom/living-room night video, hidden capability expansion increases the chance of unreviewed data access, exfiltration, and unexpected side effects far beyond user expectations.

Tp4

High
Category
MCP Tool Poisoning
Confidence
97% confidence
Finding
The documented purpose is narrow video-based rhythm anomaly analysis, but the described behavior includes local identity persistence, token acquisition, cloud history retrieval, and generic remote AI analysis that is not transparently tied to the claimed function. This mismatch is dangerous because users and reviewers may approve a privacy-sensitive health-monitoring skill without realizing it creates accounts, stores identifiers, and sends data to remote services under broader capabilities than advertised.

Context-Inappropriate Capability

Medium
Confidence
89% confidence
Finding
Cloud-based historical report querying and report-link retrieval extend the skill from one-shot local analysis into longitudinal cloud data access. For a skill processing intimate night-time in-home monitoring data, expanding to cloud history and report distribution increases exposure of sensitive behavioral patterns and creates additional attack surface for unauthorized access or overcollection.

Context-Inappropriate Capability

High
Confidence
96% confidence
Finding
Automatically creating or reusing a local default user identity is an identity-management behavior that is not required for simple video analysis and can silently bind sensitive health/behavioral reports to persistent identities. This is especially risky because the subject is a person living alone in a private bedroom/living room context, making unauthorized profiling, cross-session tracking, and mistaken identity association more harmful.

Description-Behavior Mismatch

Medium
Confidence
91% confidence
Finding
The code accepts arbitrary http/https URLs as video input even though the skill description frames this as analysis of footage from a fixed in-home camera. That broadens the trust boundary and can enable misuse such as analyzing unapproved remote content, indirect access to internal-only resources if the backend fetches URLs server-side, or collection of data outside the declared purpose.

Description-Behavior Mismatch

Medium
Confidence
86% confidence
Finding
The implementation exposes a generic `skill.get_output_analysis(input_path)` flow with no code-level enforcement that inputs are limited to the declared use case of night-time, fixed-camera, living-alone rhythm anomaly detection. This mismatch can enable broader video analysis than users or platform reviewers would expect, increasing the risk of unauthorized surveillance or repurposing of sensitive bedroom/living-room footage.

Context-Inappropriate Capability

Medium
Confidence
93% confidence
Finding
The skill accepts arbitrary remote video URLs via `--url` and forwards them for analysis without origin restrictions or purpose validation. In a skill intended for fixed in-home camera analysis, this expands the trust boundary and can be abused to analyze third-party or internet-sourced videos, potentially enabling privacy violations, SSRF-like backend fetching behavior, or processing of untrusted content outside the stated purpose.

Description-Behavior Mismatch

High
Confidence
98% confidence
Finding
This utility performs remote account bootstrap, token recovery, authenticated API calls, and header injection that are unrelated to the declared purpose of local night-video rhythm anomaly detection. In a privacy-sensitive home-camera skill, unexpected network transmission and credential handling materially increase the risk of covert data exfiltration, user tracking, and unauthorized remote service coupling.

Context-Inappropriate Capability

High
Confidence
98% confidence
Finding
The code resolves identity from local files, cached database records, or generates a new default user identifier and persists it, then uses that identity for downstream service access. For a skill whose stated function is local camera-based sleep/activity anomaly analysis, silently creating or recovering identities enables undisclosed user correlation across runs and can bind a private home-monitoring workflow to remote accounts without informed consent.

Context-Inappropriate Capability

Medium
Confidence
84% confidence
Finding
The workspace-detection logic probes environment variables, file paths, and directory structure to locate agent roots and create data/skills directories. While not overtly malicious, this exceeds the stated needs of a fixed-camera anomaly detector and expands filesystem awareness and write scope, which can facilitate persistence, cross-workspace interference, or collection of broader runtime context if reused unsafely.

Vague Triggers

Medium
Confidence
84% confidence
Finding
Overly broad trigger phrases can cause the skill to auto-invoke on generic requests and perform cloud history queries or other sensitive actions without sufficiently specific user intent. In a privacy-sensitive monitoring skill, accidental invocation can expose private historical reports or initiate data processing the user did not clearly request.

Missing User Warnings

High
Confidence
95% confidence
Finding
The skill handles highly sensitive bedroom/living-room night video and appears to rely on cloud/API processing, yet the primary description does not clearly warn users that intimate in-home footage may be uploaded to remote services. Lack of prominent disclosure undermines informed consent and can lead to severe privacy harm if users unknowingly transmit private behavioral and health-related data.

Missing User Warnings

Medium
Confidence
92% confidence
Finding
The document specifies continuous overnight monitoring of a person living alone, collection of highly sensitive behavioral data, and report export capabilities, but does not define privacy safeguards such as informed consent, retention limits, access controls, family-notification rules, or secure handling of recordings and derived results. In this context, omission of those controls can lead to surveillance misuse, unauthorized disclosure of intimate household activity patterns, and secondary harms to vulnerable individuals.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The script silently resolves and initializes an internal user identity via OpenIdUtil.resolve_current_open_id(...) even though the parameter is hidden from help output and the comment explicitly states the user is not required to input it. In a privacy-sensitive monitoring skill that analyzes in-home night video and can list historical analysis records, undisclosed identity binding can cause data to be associated with or retrieved for an unintended account without meaningful user awareness or consent.

Missing User Warnings

Medium
Confidence
88% confidence
Finding
The skill reads the entire local file and uploads it for remote analysis without any visible confirmation, notice, or minimization step in this code path. Because the skill processes bedroom/living-room night video for a person living alone, the content is highly sensitive and silent upload materially increases privacy and consent risk.

External Transmission

Medium
Category
Data Exfiltration
Content
"source": ConstantEnum.DEFAULT__SKILL_HUB_NAME
            }
            try:
                _response = requests.post(_url, json=_data)
                if _response.status_code == 200:
                    _response_json = _response.json()
                    if _response_json and _response_json.get("success"):
Confidence
97% confidence
Finding
requests.post(_url, json=

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

Detected: suspicious.install_untrusted_source

Install source points to URL shortener or raw IP.

Warn
Code
suspicious.install_untrusted_source
Location
skills/smyx_common/scripts/config-dev.yaml:2