Security audit

live-strea组合包

Security checks for vulnerabilities and agentic risk

Overview

This bundle is not clearly malicious, but it asks for broad development powers while its documentation makes unsupported capability and API-key claims without enough scope or privacy detail.

Review before installing. Use this only in a sandboxed or low-risk workspace unless you can verify the referenced member skills separately. Do not provide real API keys, private stream URLs, proprietary source code, or sensitive templates until the publisher documents which services receive data and what commands or files the bundle may modify.

SkillSpector

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
Behavioral ASTexec() Call, eval() Call, Dynamic Import
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands

Findings (6)

Intent-Code Divergence

Medium

Confidence: 91% confidence
Finding: The documentation claims real-time live-stream analysis, audience metrics, and related capabilities that are not substantiated by the bundle manifest itself. This is dangerous because users may trust the bundle with operational or sensitive workflows under false assumptions, leading to unsafe deployment decisions or exposure of data to tools that do not actually provide the advertised controls.

Intent-Code Divergence

Medium

Confidence: 95% confidence
Finding: The input/output section presents concrete behavior, structured outputs, and analytics results as if they are guaranteed by the artifact, but the bundle file does not define any executable logic or validated interface for them. Such fabricated interface documentation can cause downstream systems or users to rely on nonexistent processing, creating integrity and security risks when sensitive files, stream URLs, or code paths are supplied to an unverified toolchain.

Intent-Code Divergence

Low

Confidence: 88% confidence
Finding: The setup guidance tells users to configure API keys and establish API connections even though no corresponding API integration is defined in this artifact. This can prompt users to provision secrets unnecessarily, increasing the chance of credential mishandling, accidental disclosure, or insecure environment configuration.

Missing User Warnings

Medium

Confidence: 84% confidence
Finding: The bundle advertises exec and write capabilities as part of normal workflow guidance without a clear warning that these tools can execute commands and modify the local system or files. In a Development-category bundle, users may be more likely to run it in privileged environments, making understated system-modifying behavior more dangerous.

Missing User Warnings

Medium

Confidence: 90% confidence
Finding: The documentation instructs users to configure API keys and call APIs without explicitly warning that data, prompts, file contents, or metadata may be transmitted to external services. This is dangerous because users may unknowingly expose proprietary code, stream data, or credentials during setup and operation.

Vague Triggers

Medium

Confidence: 88% confidence
Finding: The manifest describes a broad, end-to-end workflow across multiple bundled skills with powerful capabilities including exec, write, glob, and grep, but it does not define activation boundaries, preconditions, or exclusion rules. In a bundled skill context, this increases the chance that an agent will invoke high-risk tools too broadly or in unintended situations, leading to over-privileged execution, unsafe file modification, or command use beyond the user’s actual request.

Static analysis

No suspicious patterns detected.