ClawHub
Back to skill

Security audit

GIF Multi

Security checks across malware telemetry and agentic risk

Overview

The skill does what it advertises: it searches Giphy, converts GIFs, and sends them, with some disclosed persistence and network behavior users should understand.

Install only if you are comfortable with the agent sending GIF media in your conversations, sending GIF search queries to Giphy, and storing short-lived generated media plus persistent channel and mode preferences. Use on_request mode if you do not want spontaneous GIFs.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (4)

Vague Triggers

Medium
Confidence
80% confidence
Finding
The skill allows persistent mode changes based on conversational phrases such as 'stop sending GIFs without asking' or 'feel free to send GIFs naturally.' Ambiguous natural-language state changes can be triggered unintentionally, manipulated by quoted text, or induced by third-party content, causing the agent to alter future behavior without clear authorization.

Missing User Warnings

Medium
Confidence
83% confidence
Finding
The skill documents persistent writes to config.json, caching under .gif_cache/, and automatic deletion of orphaned files, but it does not clearly warn users that preferences and channel state are stored across sessions and that files may be removed automatically. Insufficient disclosure around persistence and deletion can surprise users, complicate auditing, and create integrity or privacy issues if cached media or settings are retained longer than expected.

Missing User Warnings

Medium
Confidence
74% confidence
Finding
The skill writes downloaded and converted media into a shared cache directory and later deletes files based on age, without surfacing this behavior to users. In a multi-skill or shared-workspace environment, this can expose user activity metadata or unintentionally delete other files if directory ownership and boundaries are not tightly controlled.

Missing User Warnings

Medium
Confidence
78% confidence
Finding
The skill sends user queries to Giphy and downloads remote media without an explicit user-facing disclosure at the point of use. In an agent environment, silent external transmission can leak sensitive prompts or interests to third parties, making the context more dangerous than a normal standalone CLI tool.

VirusTotal

59/59 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.