Security audit
Security checks for vulnerabilities and agentic risk
This skill should be reviewed carefully because it handles sensitive child/home video through cloud APIs while silently creating or reusing identity tokens and claiming real-time alerting that the bundled code does not implement.
Install only if you are comfortable sending child/home surveillance videos or video URLs to the configured LifeEmergence cloud service and having report history tied to a silent local/cloud identity. Do not rely on this package as a real-time safety alarm unless the publisher separately demonstrates live camera monitoring, alert delivery, retention controls, and deletion/consent handling.
The skill appears to perform authenticated external requests, identity/token resolution, and local file handling beyond the narrow, safety-oriented description presented to users. In the context of children's videos and household surveillance, undisclosed remote processing and identity linkage materially elevate privacy and security risk.
The skill appears to perform authenticated external requests, identity/token resolution, and local file handling beyond the narrow, safety-oriented description presented to users. In the context of children's videos and household surveillance, undisclosed remote processing and identity linkage materially elevate privacy and security risk.
The skill appears to perform authenticated external requests, identity/token resolution, and local file handling beyond the narrow, safety-oriented description presented to users. In the context of children's videos and household surveillance, undisclosed remote processing and identity linkage materially elevate privacy and security risk.
The skill appears to perform authenticated external requests, identity/token resolution, and local file handling beyond the narrow, safety-oriented description presented to users. In the context of children's videos and household surveillance, undisclosed remote processing and identity linkage materially elevate privacy and security risk.
The skill appears to perform authenticated external requests, identity/token resolution, and local file handling beyond the narrow, safety-oriented description presented to users. In the context of children's videos and household surveillance, undisclosed remote processing and identity linkage materially elevate privacy and security risk.
The skill appears to perform authenticated external requests, identity/token resolution, and local file handling beyond the narrow, safety-oriented description presented to users. In the context of children's videos and household surveillance, undisclosed remote processing and identity linkage materially elevate privacy and security risk.
The skill appears to perform authenticated external requests, identity/token resolution, and local file handling beyond the narrow, safety-oriented description presented to users. In the context of children's videos and household surveillance, undisclosed remote processing and identity linkage materially elevate privacy and security risk.
The skill appears to perform authenticated external requests, identity/token resolution, and local file handling beyond the narrow, safety-oriented description presented to users. In the context of children's videos and household surveillance, undisclosed remote processing and identity linkage materially elevate privacy and security risk.
The skill appears to perform authenticated external requests, identity/token resolution, and local file handling beyond the narrow, safety-oriented description presented to users. In the context of children's videos and household surveillance, undisclosed remote processing and identity linkage materially elevate privacy and security risk.
The skill appears to perform authenticated external requests, identity/token resolution, and local file handling beyond the narrow, safety-oriented description presented to users. In the context of children's videos and household surveillance, undisclosed remote processing and identity linkage materially elevate privacy and security risk.
The skill appears to perform authenticated external requests, identity/token resolution, and local file handling beyond the narrow, safety-oriented description presented to users. In the context of children's videos and household surveillance, undisclosed remote processing and identity linkage materially elevate privacy and security risk.
The file defines persistent user-account storage and lookup behavior that is not justified by the stated purpose of detecting dangerous child window-climbing behavior from camera feeds. In a safety-monitoring skill, collecting and managing unrelated user identities increases the attack surface and creates unnecessary privacy and data-protection risk if the local database is exposed or misused.
The User model stores sensitive personal data and authentication-like secrets, including email, birthday, token, and open_token, without any evident need tied to child climbing detection. For this skill context, such retention is especially concerning because the functionality is safety monitoring, not account or identity management, so the data collection appears excessive and raises confidentiality and compliance risks.
The utility layer provisions users and performs remote login/token acquisition through /sys/phoneLogin, which is unrelated to the declared child-window-climbing video safety function. This creates hidden account creation and outbound identity flows that can enroll users, persist tokens, and expand the system's privileges without clear necessity or informed consent.
The skill documentation indicates capabilities consistent with shell execution, network access, file reads/writes, and environment access, but it does not declare any explicit tool scope or permissions boundaries. In an agent setting, this increases the risk of unintended or over-privileged execution, especially because the skill instructs command execution and cloud interaction while handling sensitive child video/report data.
The manifest and docs claim real-time fixed-camera detection with immediate alerting, but the described implementation is a batch analysis/reporting workflow. In a child fall-prevention context, this is dangerous because latency and lack of continuous monitoring can defeat the core safety expectation users rely on.
The skill discusses on-site snapshots and historical reports but does not clearly warn users that these may be uploaded to, stored in, or retrievable from cloud services. Because the content involves minors and home surveillance footage, insufficient disclosure materially increases privacy and compliance risk.
The default activation condition is overly broad and may trigger analysis on any uploaded window/balcony video without clear boundaries or user confirmation. In this context, that can lead to accidental processing of sensitive household and child footage, potentially sending it to remote services without sufficiently explicit intent.
The script exposes a history-listing capability via --list and show_analyze_list(open_id) that is not reflected in the skill's stated purpose of real-time child window-climbing detection. Because this feature retrieves prior analysis records tied to a user identity, it expands the data-access surface and can enable unauthorized disclosure of surveillance-derived history if access control is weak elsewhere.
This code presents the command description, argument help text, and runtime status/error messages only in Chinese, which imposes a specific language on all users. The file provides no option to choose another language or locale, and there is no documented justification that the skill is region-specific.
The code initializes an internal identity through OpenIdUtil.resolve_current_open_id(...) and then uses ConstantEnum.CURRENT__OPEN_ID for listing analysis records, even though identity handling is not necessary for simple local/video danger detection. In a home-surveillance context involving children, hidden identity linkage increases privacy sensitivity and creates risk of user tracking or cross-user data access if the identifier is misused.
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
Natural-language policy violations apply to all file types, including code string literals. This file returns multiple user-facing messages entirely in Chinese, such as report headings, error text, and status messages, with no visible opt-in, language selection, or documented justification that the skill is intended only for a Chinese-language context.
The skill reads local video files or forwards remote video URLs to an external analysis service without any visible consent prompt, privacy notice, or user-facing disclosure in this code path. Because the content may contain children, homes, and sensitive interior imagery, silent upload/processing increases privacy and compliance risk if invoked without explicit user understanding.
Detected: suspicious.install_untrusted_source