Security audit

UAPI 生成摸摸头GIF (QQ号) 接口

Security checks for vulnerabilities and agentic risk

Overview

This skill is a small documented wrapper for one public GIF-generation API, though its keywords are broader than the API it actually supports.

Install only if you want a skill for generating a motou GIF from a QQ number through uapis.cn. Avoid relying on it for compression, base64 conversion, SVG conversion, or NSFW detection despite those keywords appearing in the artifact; those labels should be removed by the publisher.

A.I.G

Vulnerability Patterns

Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs

SkillSpector

Vulnerability Patterns

Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access

Findings (6)

Description-Behavior Mismatch

Medium

Confidence: 98% confidence
Finding: The manifest description advertises capabilities far beyond the only documented endpoint, including image compression, image/base64 conversion, and other image tasks unrelated to /image/motou. This can cause the orchestrator to invoke the skill for requests it cannot safely satisfy, leading to incorrect API usage, unintended data handling, or misleading task routing.

Vague Triggers

Medium

Confidence: 96% confidence
Finding: Broad English trigger phrases like 'image compression', 'image to base64', and 'base64 to image' make accidental invocation likely for many unrelated image workflows. This weak query-to-skill precision and can route user data or requests into an unsuitable skill path, undermining reliability and potentially exposing inputs to unnecessary third-party processing.

Vague Triggers

Medium

Confidence: 95% confidence
Finding: The keyword list includes multiple unrelated capabilities without any exclusion rules, so a planner may match on generic image terms and invoke this skill despite it only supporting /image/motou. Because the skill calls an external API, misrouting can also send user-supplied identifiers or content to a third party when no such call was necessary.

Intent-Code Divergence

Medium

Confidence: 95% confidence
Finding: The keyword section contradicts the skill's declared single-purpose scope by listing unrelated operations such as image compression, base64 conversion, SVG conversion, and NSFW detection. In a routing system, such contradictory documentation increases the chance the skill is selected for tasks outside its intended function, producing unsafe or invalid behavior.

Description-Behavior Mismatch

Medium

Confidence: 96% confidence
Finding: The skill explicitly states it only wraps `GET /image/motou`, but the keyword list includes unrelated capabilities such as image compression, base64 conversion, SVG conversion, and NSFW image detection. This can cause incorrect skill selection and route user requests to an endpoint that cannot satisfy them, leading to unintended data handling, user confusion, and potentially unsafe downstream behavior if agents pass inappropriate inputs or credentials while trying to fulfill mismatched tasks.

Natural-Language Policy Violations

Low

Confidence: 85% confidence
Finding: The file presents invocation keywords in Chinese and English, but the document itself assumes a Chinese-language workflow and does not explicitly state that language choice is optional. Under the policy, forced or implied language defaults without user opt-in can be a natural-language policy concern.

Static analysis

No suspicious patterns detected.