ClawHub
Back to skill

Security audit

Ephemeral Media Hosting

Security checks across malware telemetry and agentic risk

Overview

This is a disclosed setup guide for temporary media hosting, but operators should review server permissions, remote fetching, logging, and cleanup behavior before use.

Install or use this only on a server intended for public media hosting. Before deployment, adjust permissions for least privilege, document IP/source-URL logging, align the retention policy, and do not expose the remote fetch helper publicly without SSRF and abuse controls.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Findings (8)

Description-Behavior Mismatch

Medium
Confidence
84% confidence
Finding
The skill’s stated purpose is temporary media hosting, but it also includes a remote URL fetch-and-cache capability. That materially expands the trust boundary and attack surface by enabling the system to retrieve attacker-supplied URLs, which can be abused for server-side request forgery, internal network probing, or fetching untrusted content not implied by the documented purpose.

Context-Inappropriate Capability

Medium
Confidence
95% confidence
Finding
The fetch_and_cache_image function performs curl requests to arbitrary user-controlled URLs with only a superficial http/https prefix check. In a hosting environment, this is dangerous because it can be used to access internal services, cloud metadata endpoints, or other restricted resources from the server’s network position.

Intent-Code Divergence

Medium
Confidence
76% confidence
Finding
The documentation promises 7-day retention, while later monitoring logic deletes files older than 3 days under storage pressure. This mismatch is not a code-execution flaw, but it is a security/privacy integrity issue because users may make decisions based on retention guarantees that are not actually upheld.

Intent-Code Divergence

Medium
Confidence
82% confidence
Finding
The monitoring code silently deletes files older than 3 days when storage exceeds 1GB, contradicting the advertised 7-day retention period. This undermines system integrity and user expectations, especially for ephemeral sharing workflows that depend on predictable availability.

Missing User Warnings

Medium
Confidence
80% confidence
Finding
The skill describes automatic deletion and includes logging/metadata behaviors, but the high-level description does not clearly warn users that uploader IPs and source information may be retained. In a media-sharing service, hidden collection of identifying metadata creates privacy risk and can violate user expectations or policy requirements.

Missing User Warnings

Medium
Confidence
92% confidence
Finding
The upload form tells users only that files are auto-deleted after 7 days, but omits that uploader IP addresses and fetched source URLs may be stored in metadata and logs. This is a concrete privacy transparency failure because users are not informed about collection of identifying information at the point of submission.

Sudo/Root Execution

Medium
Category
Privilege Escalation
Content
# 権限設定
    sudo chown -R "$nginx_user:$nginx_user" "$media_root"
    sudo chmod -R 755 "$media_root"
    sudo chmod 775 "$media_root"/{uploads,temp,processed}
    
    # 設定ファイル作成
    cat > "$media_root/config.env" << 'EOF'
Confidence
87% confidence
Finding
sudo

Sudo/Root Execution

Medium
Category
Privilege Escalation
Content
# 権限設定
    sudo chown -R "$nginx_user:$nginx_user" "$media_root"
    sudo chmod -R 755 "$media_root"
    sudo chmod 775 "$media_root"/{uploads,temp,processed}
    
    # 設定ファイル作成
    cat > "$media_root/config.env" << 'EOF'
Confidence
87% confidence
Finding
chmod 775

VirusTotal

60/60 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.