Security audit

拼多多研究员

Security checks for vulnerabilities and agentic risk

Overview

This appears to be a legitimate Pinduoduo market-research skill that searches public web data and creates local HTML reports, with minor usability cautions but no evidence of malicious behavior.

Install this if you want a Pinduoduo-focused research assistant that may search the web, run its bundled Python report generator, and create local HTML reports. Avoid putting sensitive proprietary data into report inputs unless you are comfortable with it appearing in the generated HTML, and be aware that opening the report may load Chart.js from a CDN.

SkillSpector

Vulnerability Patterns

Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
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 (2)

Vague Triggers

Medium

Confidence: 89% confidence
Finding: The trigger rules are broad and based on natural-language intent patterns without strong exclusion criteria, which can cause the skill to activate on ambiguous requests. In a skill with WebSearch, WebFetch, Bash, and file-writing capability, over-triggering can lead to unnecessary external requests, unintended report generation, or execution of local tooling in contexts the user did not clearly request.

Natural-Language Policy Violations

Medium

Confidence: 77% confidence
Finding: Mandating Chinese as the primary language without user opt-in can override user preference and reduce transparency, especially if the surrounding conversation is in another language. While not a direct code-execution flaw, it can cause user confusion, impair informed consent, and make outputs less reviewable in multilingual environments.

Static analysis

No suspicious patterns detected.