Security audit

Public Opinion Scanner

Security checks for vulnerabilities and agentic risk

Overview

This skill openly creates Chinese-language public opinion reports from public web sources, with privacy and retention caveats users should manage.

Install only if you want Chinese-language public opinion reports from public web sources. For scans about individuals, confirm you have a legitimate reason, review the sources and sentiment labels manually, avoid collecting unnecessary personal details, and delete or restrict access to saved reports when they are no longer needed.

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
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (4)

Missing User Warnings

Medium

Confidence: 92% confidence
Finding: The skill explicitly supports scanning named individuals and generating structured reports from public sources, but it does not warn about privacy, defamation, sensitive personal data, or the need to avoid unnecessary collection of personal information. This is dangerous because aggregating, labeling, and storing reputational claims about a person can amplify inaccurate or harmful content and create privacy and compliance risks.

Vague Triggers

Medium

Confidence: 95% confidence
Finding: The trigger phrases are broad enough that the skill could activate for loosely related requests and begin collecting or synthesizing sensitive reputation data without clear user intent. In a skill that scans public opinion about companies or individuals, ambiguous activation increases the chance of unintended profiling, reputational analysis, or persistence of sensitive results.

Natural-Language Policy Violations

Low

Confidence: 96% confidence
Finding: The skill description, trigger phrases, and report template are all specified in Chinese, and the output is defined as a Chinese-language markdown report without offering a language choice. This imposes a locale/language preference on users rather than allowing opt-in or documenting that the skill is intentionally region-specific.

Missing User Warnings

Low

Confidence: 84% confidence
Finding: The skill states that reports are saved to a local path but does not warn users that potentially sensitive reputation reports will persist on disk. Even when sourced from public data, these compiled reports may contain personal or reputationally damaging summaries, and silent persistence increases the risk of unintended access, retention, or redistribution.

Static analysis

No suspicious patterns detected.