Security audit

视频号爆款拆解流水线

Security checks for vulnerabilities and agentic risk

Overview

The skill is a disclosed WeChat Channels video analysis pipeline that downloads user-provided videos, transcribes them locally, and writes reports without evidence of hidden exfiltration or destructive behavior.

Install only if you are comfortable using a logged-in Yuanbao/browser session to process WeChat Channels links and storing downloaded videos, audio, transcripts, metadata, and HTML reports on disk. Process only content you are authorized to analyze, and delete local outputs when they are no longer needed.

SkillSpector

Vulnerability Patterns

Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration

Findings (6)

Lp3

Medium

Category: MCP Least Privilege
Confidence: 96% confidence
Finding: The skill instructs execution of Python scripts, ffmpeg, and Whisper, and writes downloaded media, transcripts, and reports to disk, but it does not declare permissions for shell, network, filesystem, or environment access. This weakens user consent and platform enforcement because the skill's real capabilities exceed its stated permission surface.

Context-Inappropriate Capability

Low

Confidence: 84% confidence
Finding: The script dynamically locates another skill via environment variables and filesystem probing, then prepends that path to sys.path and imports code from it. This creates a code-trust boundary issue: if an attacker can influence the environment or place a malicious module in a searched location, arbitrary Python code could execute when this script starts.

Vague Triggers

Medium

Confidence: 84% confidence
Finding: The trigger phrase '爆款分析' is broad enough to match ordinary marketing or content-analysis requests that may not imply consent to download third-party videos, scrape metadata, or create transcripts. Over-broad activation can cause the skill to run in contexts where the user did not intend these higher-risk actions.

Missing User Warnings

Medium

Confidence: 95% confidence
Finding: The skill clearly downloads videos, extracts audio, generates transcripts, and captures account and engagement metadata, yet the user-facing description does not prominently warn that local copies of potentially sensitive third-party content will be created. This creates privacy and compliance risk because users may not realize the scope of collection, retention, and secondary analysis.

Vague Triggers

Medium

Confidence: 84% confidence
Finding: The trigger phrase at line 33 ("视频号账号分析") is broad enough to match common user requests for general account/content analysis, which can cause this skill to activate outside its narrowly intended workflow. In this skill, unintended activation is more concerning because it chains network access, browser automation, file writes, and video downloading, increasing the blast radius of a misfire.

Vague Triggers

Medium

Confidence: 88% confidence
Finding: The trigger phrase at line 37 ("视频号爆款") is highly generic and likely to overlap with broad discussion of popular WeChat Channels content rather than an explicit request to run this automation pipeline. Because the skill performs downloads, transcription, metadata extraction, and browser-assisted access, an overly generic trigger can lead to unintended execution and unnecessary exposure to external content or session-backed actions.

Static analysis

No suspicious patterns detected.