Pet Vocal Emotion Analysis Skill | 宠物叫声情绪解析技能

• Outcome: Review
• Latest audit: Aug 23, 2026 · 3:33 PM UTC
• Version: 1.0.11

Overview

The skill can perform cloud pet vocal analysis, but it also silently creates or reuses identity records and tokens for cloud report history, so it needs careful review before installation.

Install only if you are comfortable sending pet media or URLs to the Life Emergence cloud service and allowing the skill to create or reuse a local identity with stored tokens for report history. Review and manage the workspace data files, especially smyx-api-key.txt and the smyx-common-claw.db database, if you need privacy boundaries or deletion control.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep
• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• Behavioral AST: exec() Call, eval() Call, Dynamic Import
• MCP Least Privilege: Underdeclared Capability, Wildcard Permission, Missing Permission Declaration

Findings (28)

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: 76% 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: 76% confidence
• Finding: query = query.filter(getattr(self.__model__, key) == value)

Lp3

Medium

• Category: MCP Least Privilege
• Confidence: 90% confidence
• Finding: The skill documentation describes capabilities to read/write local files, access environment-derived identity state, invoke shell commands, and call network APIs, but no permissions are declared. This creates a transparency and policy-enforcement gap: operators and users cannot accurately assess what the skill may access before execution.

Description-Behavior Mismatch

Medium

• Confidence: 92% confidence
• Finding: The stated purpose is pet vocal emotion analysis, but the skill expands to image/video analysis and cloud-hosted report retrieval. This scope expansion violates least functionality and can cause users or orchestrators to trigger broader data processing than expected, including handling media and remote report data unrelated to simple vocal translation.

Context-Inappropriate Capability

High

• Confidence: 97% confidence
• Finding: The skill silently initializes internal identity, reuses a default local user, and may create one if absent, even though identity management is not necessary for basic pet audio analysis. Hidden account linkage and local user creation can enable cross-session data association, privacy violations, and accidental access to another user's historical reports.

Context-Inappropriate Capability

Medium

• Confidence: 90% confidence
• Finding: The skill mandates cloud API access for historical report listing, which is a separate data retrieval feature beyond core pet vocal recognition. This broadens data exposure and may let a benign analysis request pivot into remote enumeration of past reports, especially when combined with hidden identity association.

Description-Behavior Mismatch

High

• Confidence: 94% confidence
• Finding: The implementation materially diverges from the declared skill purpose: instead of narrowly handling pet vocal emotion analysis, it exposes a generic video-analysis and history-listing workflow. This mismatch is dangerous because users and reviewers may grant permissions or trust based on the manifest, while the code processes broader inputs and potentially accesses backend features unrelated to the stated pet-audio use case.

Description-Behavior Mismatch

Medium

• Confidence: 88% confidence
• Finding: Accepting arbitrary network video URLs broadens the skill far beyond its advertised pet voiceprint purpose and can turn it into a general remote-content analysis proxy. In skill ecosystems, this increases the risk of hidden data flows, misuse of backend processing, and user deception about what content is being handled.

Description-Behavior Mismatch

Medium

• Confidence: 93% confidence
• Finding: This file exposes generic CRUD wrappers and arbitrary HTTP helpers that can send requests to caller-supplied URLs, which is far broader than the manifest’s stated purpose of pet vocal emotion analysis. In a skill context, this expands the attack surface and enables the skill or dependent code to interact with unrelated backend services or external endpoints without any purpose limitation visible here.

Context-Inappropriate Capability

Medium

• Confidence: 95% confidence
• Finding: The http_post/http_put/http_get/http_delete methods provide unrestricted remote request capability to arbitrary URLs, effectively making this skill component a generic network client. For a pet emotion analysis skill, this is unjustified functionality that could be abused for data exfiltration, hidden secondary service access, or calling attacker-controlled infrastructure if untrusted input reaches these methods.

Description-Behavior Mismatch

High

• Confidence: 97% confidence
• Finding: This file implements a reusable local user/account persistence layer, including user lookup and lifecycle management, which is unrelated to pet vocal emotion analysis. In this skill context, hidden identity storage materially expands data collection and attack surface without a clear functional need, making the capability suspicious and more dangerous than it would be in an authentication-focused skill.

Context-Inappropriate Capability

High

• Confidence: 99% confidence
• Finding: The User model stores identity and credential-like data including username, email, token, and open_token, none of which are justified by the stated pet emotion analysis purpose. In this context, collecting and persisting tokens is particularly risky because compromise of the local database could expose reusable authentication material unrelated to the advertised functionality.

Context-Inappropriate Capability

Medium

• Confidence: 89% confidence
• Finding: The skill exposes a generic `ai_chat`/agent orchestration entry point that is unrelated to the stated pet vocal emotion analysis purpose. Even though the subprocess invocation is currently commented out, this creates an unnecessary capability surface and a ready-made hook for future external command or LLM-agent execution, which increases the risk of misuse, prompt exfiltration, or privilege creep in a skill that should be narrowly scoped.

Description-Behavior Mismatch

High

• Confidence: 98% confidence
• Finding: This utility layer performs broad identity bootstrap, token handling, local persistence, and remote API calls that are not justified by a pet vocal emotion analysis skill. That mismatch materially increases risk because installing or invoking the skill can trigger unrelated account and network operations, expanding the attack surface and creating opportunities for covert data use or platform abuse.

Context-Inappropriate Capability

High

• Confidence: 97% confidence
• Finding: The code reads identity values from workspace files, reuses local database identities, and creates persistent default open-id records without a clear user-facing reason tied to pet emotion analysis. Silent identity creation and reuse can let the skill act on behalf of a user or pseudo-user without informed consent, enabling tracking, unintended account linkage, or unauthorized downstream requests.

Context-Inappropriate Capability

High

• Confidence: 99% confidence
• Finding: The helper auto-registers or logs in a user against a remote health endpoint using phoneLogin-style semantics with silent/register flags, which is unrelated to analyzing cat or dog vocalizations. This can create or access remote accounts without clear authorization, potentially transmitting identifiers off-device and enabling unauthorized service consumption or user impersonation within that platform.

Context-Inappropriate Capability

Medium

• Confidence: 91% confidence
• Finding: The code performs workspace detection plus creation of agent data and skills directories, capabilities that are broader than needed for pet sound interpretation. In a skill context, filesystem and workspace discovery can facilitate unintended persistence, cross-skill interaction, or placement of files in sensitive agent-controlled locations.

Vague Triggers

Medium

• Confidence: 88% confidence
• Finding: The history-report trigger phrases are broad enough to overlap with ordinary conversational requests, increasing the chance of unintended cloud queries. In context, this is more dangerous because triggering report lookup can expose prior report metadata tied to hidden identity state rather than only processing the current media item.

Missing User Warnings

Medium

• Confidence: 94% confidence
• Finding: The skill instructs automatic saving of uploaded attachments to local storage without a user-facing warning or retention controls. This creates unnecessary local persistence of potentially sensitive media and increases exposure to data leakage, reuse by other tasks, or forensic recovery.

Missing User Warnings

Medium

• Confidence: 84% confidence
• Finding: The skill reads arbitrary local file contents and forwards them to an external analysis service, while this file provides no user-facing disclosure, consent prompt, or narrowing to pet-audio-specific inputs. In the context of a mislabeled pet emotion skill, that increases the risk users may unknowingly upload unrelated local files, causing unintended data exfiltration to a remote service.

Missing User Warnings

Medium

• Confidence: 91% confidence
• Finding: The tool initiates analysis of local file paths or remote URLs through a backend-facing skill method without clearly warning the user that content or references may be transmitted off-device. This creates a privacy and consent problem, especially when users may believe analysis is local or limited to benign pet-sound processing.

Missing User Warnings

Medium

• Confidence: 78% confidence
• Finding: The helper performs outbound HTTP requests to retrieve a download URL with no visible disclosure, consent handling, or indication in this file that users are informed about network transmission. While network access may be operationally necessary, the lack of transparency is risky in a user-facing skill because uploaded identifiers or related metadata may be sent to remote services without clear notice.

Missing User Warnings

Medium

• Confidence: 87% confidence
• Finding: The function trusts an environment-derived workspace path and later uses that context for directory creation, without user disclosure or clear validation. In agent environments, environment variables can be influenced by the surrounding runtime, so this may cause data or files to be written into unintended locations, increasing persistence and boundary-crossing risk.

Missing User Warnings

Medium

• Confidence: 95% confidence
• Finding: The code silently persists identity data locally by creating a default username and storing user records for later reuse. Even if intended as convenience behavior, undisclosed identifier retention can enable tracking across sessions and remote account coupling that users would not reasonably expect from a pet emotion analysis feature.

Missing User Warnings

Medium

• Confidence: 96% confidence
• Finding: This code assembles and transmits usernames, open IDs, tokens, API keys, and related metadata in HTTP requests without clear non-debug disclosure to the user. The danger is amplified by the skill mismatch: a pet-audio skill should not silently handle platform identity and auth material unless that behavior is essential and transparently documented.

VirusTotal

VirusTotal findings are pending for this skill version.

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