ClawHub
Back to skill

Security audit

发布快手短视频_无限

Security checks for vulnerabilities and agentic risk

Overview

This skill is purpose-built to publish Kuaishou videos, but it tells the agent to complete public posting and change download permissions without a clear final user confirmation checkpoint.

Install only if you are comfortable with an agent operating the Kuaishou creator page and submitting videos. Before use, require the agent to summarize the exact video file, title/description, tags, cover, AI-content declaration, and download setting, then wait for your explicit approval before clicking publish.

A.I.G

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (2)

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The skill automates posting to Kuaishou and modifying a privacy-relevant setting ('允许下载此作品') but never clearly warns that it will make a public-facing platform change. A user invoking 'upload/post Kuaishou video' may expect assistance, yet the skill is designed to carry the workflow through to publication and alter download permissions without an explicit confirmation or warning about these consequences.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The final step instructs the agent to click '发布' to complete the workflow, but it does not explicitly state that this action is irreversible or that it will submit the video live/on-platform. In an autonomous browser automation context, this increases the risk of accidental posting because the agent is also told to solve problems itself and minimize asking the user.

Content

No source excerpt is available for this finding.