Security audit
Security checks for vulnerabilities and agentic risk
This skill is a cloud-backed pet video analysis tool, but it also silently creates or reuses identities, stores auth tokens locally, and does not implement the advertised soothing-device trigger behavior.
Review this carefully before installing. It sends media or media URLs to a remote service, associates reports with an automatically chosen identity, and may store authentication tokens locally. Do not use it for sensitive household footage unless you are comfortable with that cloud processing and local credential persistence.
This finding points to a combination of authenticated external API calls, local file read/write, user identity resolution, database access, and token storage that are substantially broader than the advertised function. In context, that is dangerous because a pet-analysis skill would not reasonably require such expansive hidden capabilities, so users and hosting agents may grant trust under false assumptions.
This finding points to a combination of authenticated external API calls, local file read/write, user identity resolution, database access, and token storage that are substantially broader than the advertised function. In context, that is dangerous because a pet-analysis skill would not reasonably require such expansive hidden capabilities, so users and hosting agents may grant trust under false assumptions.
This finding points to a combination of authenticated external API calls, local file read/write, user identity resolution, database access, and token storage that are substantially broader than the advertised function. In context, that is dangerous because a pet-analysis skill would not reasonably require such expansive hidden capabilities, so users and hosting agents may grant trust under false assumptions.
This finding points to a combination of authenticated external API calls, local file read/write, user identity resolution, database access, and token storage that are substantially broader than the advertised function. In context, that is dangerous because a pet-analysis skill would not reasonably require such expansive hidden capabilities, so users and hosting agents may grant trust under false assumptions.
This finding points to a combination of authenticated external API calls, local file read/write, user identity resolution, database access, and token storage that are substantially broader than the advertised function. In context, that is dangerous because a pet-analysis skill would not reasonably require such expansive hidden capabilities, so users and hosting agents may grant trust under false assumptions.
This finding points to a combination of authenticated external API calls, local file read/write, user identity resolution, database access, and token storage that are substantially broader than the advertised function. In context, that is dangerous because a pet-analysis skill would not reasonably require such expansive hidden capabilities, so users and hosting agents may grant trust under false assumptions.
This finding points to a combination of authenticated external API calls, local file read/write, user identity resolution, database access, and token storage that are substantially broader than the advertised function. In context, that is dangerous because a pet-analysis skill would not reasonably require such expansive hidden capabilities, so users and hosting agents may grant trust under false assumptions.
This finding points to a combination of authenticated external API calls, local file read/write, user identity resolution, database access, and token storage that are substantially broader than the advertised function. In context, that is dangerous because a pet-analysis skill would not reasonably require such expansive hidden capabilities, so users and hosting agents may grant trust under false assumptions.
This finding points to a combination of authenticated external API calls, local file read/write, user identity resolution, database access, and token storage that are substantially broader than the advertised function. In context, that is dangerous because a pet-analysis skill would not reasonably require such expansive hidden capabilities, so users and hosting agents may grant trust under false assumptions.
This finding points to a combination of authenticated external API calls, local file read/write, user identity resolution, database access, and token storage that are substantially broader than the advertised function. In context, that is dangerous because a pet-analysis skill would not reasonably require such expansive hidden capabilities, so users and hosting agents may grant trust under false assumptions.
This finding points to a combination of authenticated external API calls, local file read/write, user identity resolution, database access, and token storage that are substantially broader than the advertised function. In context, that is dangerous because a pet-analysis skill would not reasonably require such expansive hidden capabilities, so users and hosting agents may grant trust under false assumptions.
This finding points to a combination of authenticated external API calls, local file read/write, user identity resolution, database access, and token storage that are substantially broader than the advertised function. In context, that is dangerous because a pet-analysis skill would not reasonably require such expansive hidden capabilities, so users and hosting agents may grant trust under false assumptions.
This finding points to a combination of authenticated external API calls, local file read/write, user identity resolution, database access, and token storage that are substantially broader than the advertised function. In context, that is dangerous because a pet-analysis skill would not reasonably require such expansive hidden capabilities, so users and hosting agents may grant trust under false assumptions.
This finding points to a combination of authenticated external API calls, local file read/write, user identity resolution, database access, and token storage that are substantially broader than the advertised function. In context, that is dangerous because a pet-analysis skill would not reasonably require such expansive hidden capabilities, so users and hosting agents may grant trust under false assumptions.
YARA rule matched a hack tool or exploit indicator (offensive tools, reconnaissance, privilege escalation, or exploit frameworks).
---
name: "pet-calming-trigger-analysis"
description: "Automatically triggers soothing mechanisms (playing relaxing sounds, activating laser toys) when pet anxiety, howling, or prolonged loneliness is detected; a smart companion for pet care. | 宠物安抚触发技能，检测到宠物焦虑、嚎叫、长时间孤独时，自动触发安抚机制（播放舒缓音效、开启激光逗宠），智能宠物陪伴好帮手"
version: "1.0.22"
license: "MIT-0"
---
# 🧸 Pet Soothing Trigger Analysis Skill | 宠物安抚触发分析技能
> **智能分析中枢** · 图片/视频智能分析 · 结构化报告 · 历史报告云端查�
The implementation materially diverges from the declared pet-calming purpose and instead provides a generic video/file analysis and report export capability. This kind of hidden or mismatched functionality is dangerous because it can be used to process arbitrary content and exfiltrate analysis results under the guise of a benign pet-care skill, defeating user and platform trust boundaries.
The code implements generic video analysis and history retrieval via skill.get_output_analysis and skill.get_output_analysis_list, which is materially inconsistent with the declared pet-calming trigger functionality. This kind of capability mismatch is dangerous because it can conceal unrelated data-processing behavior behind an innocuous manifest, enabling covert collection or exfiltration of user video content and metadata under false pretenses.
This file exposes generic CRUD wrappers and arbitrary HTTP methods that accept caller-supplied URLs, making the skill capable of interacting with unrelated external services far beyond its stated pet-calming purpose. In a skill context, this broad network surface can be abused for unauthorized data exfiltration, command-and-control style traffic, or reaching internal services if untrusted inputs can influence the target URL.
The User model stores identity and authentication-related data such as username, realname, email, token, and open_token, which is difficult to justify for a pet-soothing trigger skill. If this component is exposed or reused broadly, it creates unnecessary risk of credential/token leakage and privacy violations far beyond the advertised functionality.
This utility implements generic authenticated HTTP access, token reuse, remote phone-login, and automatic user registration logic that is unrelated to the stated pet-soothing functionality. In skill context, that greatly increases risk because the skill can perform broad remote actions under silently obtained credentials, creating opportunities for unauthorized data transmission or account activity.
The skill declares no explicit tool scope even though the manifest and workflow instruct use of shell execution, network access, and local file handling. Missing least-privilege declarations can cause the agent to run with broader capabilities than users expect, increasing the blast radius if the skill or its dependencies are abused.
The default trigger activates on broad pet-video analysis requests, which can cause the skill to run in situations where the user did not specifically consent to cloud analysis, local file saving, or history-linked processing. Overbroad auto-invocation is risky because it increases unintended exposure of uploaded media and expands opportunities for misuse of privileged actions.
The skill states that uploaded attachments are automatically saved locally, but it does not present this as a clear user warning in the high-level description or consent flow. Silent local persistence of user media creates privacy and retention risks, especially for home-monitoring videos that may contain sensitive household information.
The skill mandates cloud API queries for history retrieval but does not prominently warn users that their requests are tied to cloud-side records and internal identity association. This is dangerous because users may believe analysis is local while the skill actually transmits or retrieves potentially sensitive pet/home-monitoring metadata from a remote service.
The manifest describes a skill that detects pet anxiety/howling/loneliness and then automatically activates responses such as relaxing sounds or laser toys. In this file, the implemented behavior is limited to calling analysis/listing methods (get_output_analysis and get_output_analysis_list) and printing/saving results, with no code that initiates any soothing mechanism.
Detected: suspicious.install_untrusted_source