ClawHub
Back to skill

Security audit

Bilibili All In One Free

Security checks across malware telemetry and agentic risk

Overview

This Bilibili skill is mostly aligned with its stated purpose, but it asks for broad command execution and an unclear API key while also downloading files locally without enough user-control guidance.

Review this skill before installing. Only provide an API key if you know exactly which service requires it, choose download locations explicitly, avoid overwriting existing files, and expect shell commands, package installs, network requests, and ffmpeg processing when using download features.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • 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
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Findings (4)

Intent-Code Divergence

Medium
Confidence
94% confidence
Finding
The documentation makes contradictory trust claims: it repeatedly says the skill requires no credentials because it uses public Bilibili APIs, yet it also instructs users to export an API key. This can mislead operators into provisioning unnecessary secrets, increasing the chance of secret exposure or accidental reuse of unrelated API credentials in an execution-capable skill.

Intent-Code Divergence

Medium
Confidence
88% confidence
Finding
The feature matrix is internally inconsistent: the free version claims support for 13 ranking categories, while the upgrade section says full 13-category ranking access is a paid feature. These contradictions erode operator trust and can cause users or agents to invoke unsupported functionality, creating unsafe fallback behavior or unexpected tool execution patterns.

Vague Triggers

Medium
Confidence
85% confidence
Finding
The activation guidance is overly broad, telling the agent to use the skill whenever a task relates to its capabilities without defining boundaries, eligibility checks, or disallowed contexts. In an exec-enabled skill, vague invocation rules increase the risk of unintended activation, overreach, or misuse on requests that did not clearly ask for network calls or downloads.

Missing User Warnings

Medium
Confidence
91% confidence
Finding
The skill advertises downloading videos as local mp4 files but does not prominently warn that this writes files to disk, may consume storage, and can modify the local environment through ffmpeg-assisted processing. In an agent context with exec access, missing disclosure about filesystem effects can lead to unanticipated local changes and reduced user control.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.