ClawHub
Back to skill

Security audit

界面设计 · UI Design

Security checks for vulnerabilities and agentic risk

Overview

This UI design guide needs review because its download instructions replace security filtering with shell commands without equivalent safeguards.

Before enabling asset downloads, require a downloader that checks destinations, redirects, and file contents. Keep downloads inside the intended project asset directory and review blocked requests instead of automatically switching tools.

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
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (1)

Asset-download guidance reframes security filtering as a tooling defect and directs the agent to an unfiltered alternative.

Medium
Category
Prompt Injection
Confidence
94% confidence
Finding

The skill explicitly replaces an SSRF/content-type-filtered download tool with shell downloads and characterizes the filtering as a tooling defect, weakening a security boundary. Attacker-controlled asset URLs or redirects could consequently reach internal services accessible from the agent environment or download unexpected content without equivalent validation. Downloading assets is legitimate for UI work, and no malicious destination is specified, but merely advising agents to avoid internal hosts does not enforce protection, particularly across redirects.

Content

No source excerpt is available for this finding.