Huo15 Douyin Video Summary

• Outcome: Review
• Latest audit: Jul 18, 2026 · 4:10 AM UTC
• Version: 1.1.0

Overview

This skill mostly matches its Douyin video download and summary purpose, but it can automatically read local browser cookies and use them for network requests without a clear opt-in.

Review before installing. Only use this skill if you are comfortable with yt-dlp reading cookies from your local Chrome, Safari, or Firefox profile to access Douyin content. Prefer running it only on clear Douyin links, and avoid using it for generic video-summary requests unless cookie use is expected.

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 (7)

subprocess module call

Medium

• Category: Dangerous Code Execution

Content

test_cmd = ["yt-dlp", "--cookies-from-browser", browser,
                        "--no-warnings", "-F", url]
            try:
                subprocess.run(test_cmd, capture_output=True, timeout=30)
                cmd.extend(["--cookies-from-browser", browser])
                print(f"  ↳ 使用 {browser} cookies")
                break

• Confidence: 86% confidence
• Finding: subprocess.run(test_cmd, capture_output=True, timeout=30)

Lp3

Medium

• Category: MCP Least Privilege
• Confidence: 92% confidence
• Finding: The skill explicitly instructs the agent to run external commands (`python3`, `yt-dlp`, `ffmpeg`) but does not declare corresponding permissions or clearly constrain shell execution. This creates a trust and review gap: users or the runtime may authorize a skill that appears low-privilege while it actually performs local command execution and network-enabled downloading.

Context-Inappropriate Capability

Medium

• Confidence: 95% confidence
• Finding: The script automatically reads cookies from local browsers, a highly sensitive capability that is not disclosed in the skill description. In an agent-skill context, hidden access to browser cookies materially increases risk because users may trigger the skill expecting media processing, not local credential use for authenticated requests.

Vague Triggers

Medium

• Confidence: 88% confidence
• Finding: The trigger phrases and description are broad enough to match generic requests such as '视频总结' or '视频转文字', which can cause the skill to activate unexpectedly outside a Douyin-specific context. Over-broad invocation increases the chance of accidental execution of download/transcoding behavior, including browser-cookie import and external network access, when the user did not intend this skill specifically.

Vague Triggers

Medium

• Confidence: 90% confidence
• Finding: The aliases include highly ambiguous names like '视频总结', '视频转文字', and '抖音下载', which can overlap with many unrelated user requests. Because this skill performs downloading and may import browser cookies, ambiguous routing makes the consequences of accidental activation more severe than for a harmless text-only skill.

Missing User Warnings

Medium

• Confidence: 94% confidence
• Finding: The skill states that browser cookies may be automatically imported from Chrome/Safari/Firefox to access Douyin content, but this sensitive behavior is not prominently disclosed up front as a privacy/security warning. Importing browser cookies exposes authentication material and browsing-session secrets to local tooling, which can materially increase account compromise risk if logs, subprocesses, or downstream tools mishandle them.

Missing User Warnings

Medium

• Confidence: 94% confidence
• Finding: The script combines automatic browser cookie access with immediate network requests and gives only a vague console message after deciding to import cookies. That means local authentication material may be transmitted to a remote service without informed consent, which is especially risky in an agent environment where execution may be one-click or automated.

VirusTotal

65/65 vendors flagged this skill as clean.

Static analysis

No suspicious patterns detected.